feat: /add-audit skill — opt-in ncl-surface audit log on the guard seam

Ships the local audit log as an install skill, scoped to the ncl command
surface: one withAudit(dispatch) composition at the definition site covers
both transports and the grant-carrying approved replay; holds are recorded
as pending events correlated by approval id (recovered from the hold row),
so --correlation returns the whole gated chain. src/audit/ installs as a
domain-free leaf — schema v1, recursive key-pattern redactor (JSON string
args parsed so inner secret keys mask), NDJSON day-files under data/audit/,
fail-open + loud emit, post-write exporter hooks, boot writability refusal,
self-contained retention prune (boot + daily, unref'd timer). `ncl audit
list` is host + global-scope only (off the group-scope allowlist, fails
closed) and errors clearly on a disabled box. Core reach-ins: the dispatch
composition and the resource-barrel import, both guarded by a shipped
AST/behavior wiring test; guard conformance stays clean at boot and in CI.
Approval-lifecycle events, OneCLI credential holds, and channel/sender
events are later increments that attach as pure adds.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

.claude/skills/add-audit/REMOVE.md

@@ -0,0 +1,52 @@
+# Remove /add-audit
+
+Reverses every change the skill made. Safe to re-run even if some pieces are
+already gone. Run from the NanoClaw project root.
+
+## 1. Delete the copied files
+
+```bash
+rm -rf src/audit
+rm -f src/cli/dispatch.audit.ts src/cli/dispatch.audit.test.ts
+rm -f src/cli/resources/audit.ts
+rm -f src/audit-wiring.test.ts
+```
+
+## 2. Revert the dispatch composition
+
+In `src/cli/dispatch.ts`, delete (not comment out) the three edits the skill
+made:
+
+1. The import line: `import { withAudit } from './dispatch.audit.js';`
+2. The composition block (comment + `export const dispatch = withAudit(dispatchInner);`)
+3. Rename the dispatcher back — change `async function dispatchInner(` to
+   `export async function dispatch(`
+
+## 3. Unregister the resource
+
+Delete the `import './audit.js';` line from `src/cli/resources/index.ts`.
+
+## 4. Remove the settings
+
+Delete the `AUDIT_ENABLED` and `AUDIT_RETENTION_DAYS` lines from `.env`.
+
+## 5. Rebuild and restart
+
+```bash
+pnpm run build
+source setup/lib/install-slug.sh
+launchctl kickstart -k gui/$(id -u)/$(launchd_label)   # macOS
+# systemctl --user restart $(systemd_unit)             # Linux
+```
+
+## Day-files
+
+`data/audit/*.ndjson` are the operator's records and are deliberately left in
+place. To purge them too:
+
+```bash
+rm -rf data/audit
+```
+
+No dependency was added, nothing under `container/` was touched, and no DB
+schema changed — there is nothing else to undo.

.claude/skills/add-audit/SKILL.md

@@ -0,0 +1,211 @@
+---
+name: add-audit
+description: Add an opt-in local audit log for the ncl command surface — every dispatch outcome on both transports (host socket and container), including scope denials, approval holds, and approved replays, written as SIEM-shaped append-only NDJSON day-files under data/audit/. Read back with `ncl audit list`; plug exporters in via registerAuditHook. Off until AUDIT_ENABLED=true.
+---
+
+# /add-audit — Local Audit Log (ncl surface)
+
+Records one canonical audit event for every command that reaches the `ncl`
+dispatcher — human over the host socket or agent over the container transport
+— including denials. Both transports converge on the exported `dispatch`, so
+one composition covers the whole surface: there is no second door.
+
+```
+ncl (host socket) ──┐
+                    ├─→ dispatch = withAudit(dispatchInner) ─→ one NDJSON line
+container cli_request ┘        │                               data/audit/<UTC-day>.ndjson
+approved replay (grant) ───────┘
+```
+
+What one event carries: `actor` (`host:<os-user>` or the agent group),
+`origin` (transport, session, channel), dotted `action` from the guard
+catalog (e.g. `groups.config.add-mcp-server`), touched/attempted `resources`,
+`outcome` (`success · failure · denied · pending`), `correlation_id` (the
+approval id on gated chains — the pending event and the approved replay share
+it, so `--correlation <id>` returns the whole chain), and redacted `details`.
+
+Scope of this increment: the ncl dispatch surface only. Approval-lifecycle
+events (request/decision as their own events), OneCLI credential holds, and
+channel/sender events are later increments — they attach to `src/audit/` as
+pure adds (new `*.audit.ts` adapters and observer subscriptions).
+
+## Steps
+
+### 0. Pre-flight (idempotency)
+
+The apply is safe to re-run; every step below is guarded. Skip to
+**Enable** if all of these already hold:
+
+- `src/audit/` exists
+- `src/cli/resources/index.ts` contains `import './audit.js';`
+- `src/cli/dispatch.ts` contains `export const dispatch = withAudit(dispatchInner);`
+
+Before editing, verify the reach-in targets still exist: `src/cli/dispatch.ts`
+must contain `export async function dispatch(` (or the already-applied
+composition), and `src/cli/resources/index.ts` must be the resource barrel (a
+list of `import './<resource>.js';` lines). If either has moved, stop and
+adapt rather than guessing.
+
+### 1. Copy the payload
+
+From the NanoClaw project root:
+
+```bash
+cp -R "${CLAUDE_SKILL_DIR}/add/src/." src/
+```
+
+What lands (mirrors of the destination paths):
+
+- `src/audit/` — the domain-free leaf: event schema (`types.ts`), env config
+  (`config.ts`), redactor, NDJSON day-file store, emit seam, post-write hooks,
+  reader, boot/maintenance wiring, vocabulary — plus its tests.
+- `src/cli/dispatch.audit.ts` — the CLI adapter: `withAudit` middleware and
+  the actor/origin/resource mapping (+ `dispatch.audit.test.ts`).
+- `src/cli/resources/audit.ts` — the read-only `ncl audit` resource.
+- `src/audit-wiring.test.ts` — goes red if either core edit below is deleted
+  or drifts.
+
+### 2. Register the resource
+
+Append to `src/cli/resources/index.ts` (skip if the line is already present):
+
+```typescript
+import './audit.js';
+```
+
+### 3. Compose the dispatch middleware
+
+This is the skill's one functional reach-in, in `src/cli/dispatch.ts`. Three
+small edits:
+
+1. Add the import (next to the other `./` imports):
+
+   ```typescript
+   import { withAudit } from './dispatch.audit.js';
+   ```
+
+2. Rename the dispatcher declaration — change
+
+   ```typescript
+   export async function dispatch(
+   ```
+
+   to
+
+   ```typescript
+   async function dispatchInner(
+   ```
+
+3. Directly after that function's closing brace (before the
+   `registerApprovalHandler('cli_command', …)` block), add:
+
+   ```typescript
+   // Audit middleware (installed by /add-audit): the exported dispatch is the
+   // wrapped function, so both transports and the approved replay below all
+   // pass the one composition.
+   export const dispatch = withAudit(dispatchInner);
+   ```
+
+The composition must live at the definition site (not at the import sites):
+the approved-replay handler in the same file calls `dispatch(...)` too, and
+only the wrapped export covers it. `src/audit-wiring.test.ts` asserts exactly
+this shape via the TypeScript AST.
+
+Loading `dispatch.audit.ts` also boots the audit log: on an enabled box it
+asserts `data/audit/` is writable (refusing to start beats a silent audit
+gap), runs the boot retention prune, and arms an unref'd maintenance timer.
+Nothing else in core changes — no `index.ts` or `host-sweep.ts` edits.
+
+### 4. Enable
+
+Add the two settings to `.env` (idempotent — overwrite or append):
+
+```bash
+grep -q '^AUDIT_ENABLED=' .env && sed -i.bak 's/^AUDIT_ENABLED=.*/AUDIT_ENABLED=true/' .env && rm -f .env.bak || echo 'AUDIT_ENABLED=true' >> .env
+grep -q '^AUDIT_RETENTION_DAYS=' .env || echo 'AUDIT_RETENTION_DAYS=90' >> .env
+```
+
+`AUDIT_ENABLED` is the master switch — off (or absent) means `emitAuditEvent`
+is a no-op and `data/audit/` is never created. `AUDIT_RETENTION_DAYS`: day
+files strictly older than the horizon are hard-deleted (unlinked) at boot and
+once per UTC day; `0` = keep forever; unset = 90.
+
+### 5. Build and test
+
+Run `build` before the tests — it's what catches a missed copy or a drifted
+import path across the whole composed tree:
+
+```bash
+pnpm run build
+pnpm exec vitest run src/audit src/cli/dispatch.audit.test.ts src/audit-wiring.test.ts
+pnpm test
+```
+
+### 6. Restart the service
+
+```bash
+source setup/lib/install-slug.sh
+launchctl kickstart -k gui/$(id -u)/$(launchd_label)   # macOS
+# systemctl --user restart $(systemd_unit)             # Linux
+```
+
+### 7. Verify (runtime smoke)
+
+```bash
+ncl groups list                     # any command — this one is now an audit event
+cat data/audit/$(date -u +%F).ndjson | tail -1
+ncl audit list --limit 5
+```
+
+The last line of the day-file is the `groups.list` event you just caused.
+
+## Reading it back
+
+```
+ncl audit list [--actor <id>] [--action <name-or-prefix>] [--resource <id-or-type>]
+               [--outcome success|failure|denied|pending|approved|rejected]
+               [--since 7d|24h|30m|ISO] [--until …] [--correlation <approval-id>]
+               [--limit N] [--format ndjson]
+```
+
+Newest first, default limit 100. `--action groups.config` matches the whole
+dotted subtree. `--format ndjson` streams the stored lines verbatim — pipe to
+a file for SIEM import. On a disabled box the command errors with
+`audit log is disabled — set AUDIT_ENABLED=true` rather than returning an
+empty list that would read as "no actions happened".
+
+The resource is deliberately **not** on the group-scope allowlist: audit
+spans agent groups, so group-scoped agents are refused before the handler
+(fails closed). Host callers and `cli_scope: global` agents only.
+
+## Redaction, failure posture, exporters
+
+- Every event's `details` passes a recursive key-pattern redactor
+  (`/(token|secret|key|password|credential|auth|bearer)/i` → `[REDACTED]`,
+  ~2 KB per-value cap) at the single emit seam.
+- Fail-open + loud: a failed append is `log.error`'d and the audited action
+  proceeds. At boot, an enabled box refuses to start if `data/audit/` isn't
+  writable.
+- In-process exporters register via `registerAuditHook` (from
+  `src/audit/index.js`) — post-write hooks that fire only after the local
+  append succeeds, so an external system can never be ahead of the source of
+  truth. Ship one as its own `/add-*` skill; declare `/add-audit` as its
+  dependency.
+
+## Troubleshooting
+
+- **Host won't start, banner names the audit directory** — `AUDIT_ENABLED=true`
+  but `data/audit/` isn't writable. Fix permissions or disable audit.
+- **`audit log is disabled — set AUDIT_ENABLED=true`** — the read-back guard;
+  enable in `.env` and restart.
+- **An agent gets "CLI access is scoped to this agent group" for `ncl audit`**
+  — by design; grant the group `cli_scope: global` only if it should read
+  cross-group history.
+- **`pnpm test` on an enabled box adds a few events to today's day-file** —
+  expected: the base dispatch tests drive real dispatches. Harmless noise;
+  the audit test suites themselves write only to temp dirs or capture buffers.
+
+## Removal
+
+See [REMOVE.md](REMOVE.md) — reverses every change; existing day-files are
+left in place (they're the operator's records).

.claude/skills/add-audit/add/src/audit-wiring.test.ts

@@ -0,0 +1,136 @@
+/**
+ * Wiring tests for the /add-audit skill's two core edits — they go red if
+ * either edit is deleted or drifts:
+ *
+ *  1. src/cli/dispatch.ts — the exported dispatch must be the composed
+ *     `withAudit(dispatchInner)` (AST check: only the definition-site
+ *     composition covers both transports AND the in-module approved replay).
+ *  2. src/cli/resources/index.ts — the audit resource must register through
+ *     the real barrel (behavior check), stay OFF the group-scope allowlist,
+ *     and leave the guard conformance walk clean.
+ *
+ * Plus the emit invariant the module's discipline rests on: emitAuditEvent
+ * appears only in src/audit/ and *.audit.ts adapter files.
+ */
+import fs from 'fs';
+import path from 'path';
+import ts from 'typescript';
+import { describe, expect, it } from 'vitest';
+
+// Same production barrels the boot conformance walk sees (mirrors
+// src/guard/conformance.test.ts).
+import './cli/commands/index.js';
+import './modules/index.js';
+import './cli/delivery-action.js';
+import './cli/dispatch.js';
+
+import { commandGuard, GROUP_SCOPE_RESOURCES, lookup } from './cli/registry.js';
+import { grantContinuationGaps } from './guard-conformance.js';
+
+describe('audit resource registration (barrel wiring)', () => {
+  it('registers audit-list through the real resource barrel', () => {
+    const cmd = lookup('audit-list');
+    expect(cmd).toBeDefined();
+    expect(cmd?.action).toBe('audit.list');
+    expect(cmd?.access).toBe('open');
+  });
+
+  it('derives a guard-catalog entry for the audit command', () => {
+    const guard = commandGuard('audit-list');
+    expect(guard.action).toBe('audit.list');
+  });
+
+  it('is NOT on the group-scope allowlist — group-scoped agents fail closed', () => {
+    expect(GROUP_SCOPE_RESOURCES.has('audit')).toBe(false);
+  });
+
+  it('leaves the boot conformance walk clean with the audit resource registered', () => {
+    expect(grantContinuationGaps()).toEqual([]);
+  });
+});
+
+describe('dispatch composition (AST wiring)', () => {
+  const source = fs.readFileSync(path.resolve('src/cli/dispatch.ts'), 'utf8');
+  const sf = ts.createSourceFile('dispatch.ts', source, ts.ScriptTarget.Latest, true);
+
+  const hasExportModifier = (node: ts.HasModifiers): boolean =>
+    (ts.getModifiers(node) ?? []).some((m) => m.kind === ts.SyntaxKind.ExportKeyword);
+
+  let importsWithAudit = false;
+  let innerDeclaredUnexported = false;
+  let exportsWrappedDispatch = false;
+  let exportsUnwrappedDispatchFn = false;
+
+  sf.forEachChild((node) => {
+    if (
+      ts.isImportDeclaration(node) &&
+      ts.isStringLiteral(node.moduleSpecifier) &&
+      node.moduleSpecifier.text === './dispatch.audit.js'
+    ) {
+      const named = node.importClause?.namedBindings;
+      if (named && ts.isNamedImports(named) && named.elements.some((e) => e.name.text === 'withAudit')) {
+        importsWithAudit = true;
+      }
+    }
+    if (ts.isFunctionDeclaration(node) && node.name?.text === 'dispatchInner') {
+      innerDeclaredUnexported = !hasExportModifier(node);
+    }
+    if (ts.isFunctionDeclaration(node) && node.name?.text === 'dispatch' && hasExportModifier(node)) {
+      exportsUnwrappedDispatchFn = true;
+    }
+    if (ts.isVariableStatement(node) && hasExportModifier(node)) {
+      for (const decl of node.declarationList.declarations) {
+        if (
+          ts.isIdentifier(decl.name) &&
+          decl.name.text === 'dispatch' &&
+          decl.initializer &&
+          ts.isCallExpression(decl.initializer) &&
+          ts.isIdentifier(decl.initializer.expression) &&
+          decl.initializer.expression.text === 'withAudit' &&
+          decl.initializer.arguments.length === 1 &&
+          ts.isIdentifier(decl.initializer.arguments[0]) &&
+          decl.initializer.arguments[0].text === 'dispatchInner'
+        ) {
+          exportsWrappedDispatch = true;
+        }
+      }
+    }
+  });
+
+  it('imports withAudit from the audit adapter', () => {
+    expect(importsWithAudit).toBe(true);
+  });
+
+  it('keeps the inner dispatcher unexported (the guarded path is the only path)', () => {
+    expect(innerDeclaredUnexported).toBe(true);
+  });
+
+  it('exports dispatch as withAudit(dispatchInner)', () => {
+    expect(exportsWrappedDispatch).toBe(true);
+    expect(exportsUnwrappedDispatchFn).toBe(false);
+  });
+});
+
+describe('emit invariant', () => {
+  it('emitAuditEvent appears only in src/audit/ and *.audit.ts files', () => {
+    const srcRoot = path.resolve('src');
+    const offenders: string[] = [];
+    const walk = (dir: string): void => {
+      for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
+        const full = path.join(dir, entry.name);
+        if (entry.isDirectory()) {
+          walk(full);
+          continue;
+        }
+        if (!entry.name.endsWith('.ts')) continue;
+        const rel = path.relative(srcRoot, full).split(path.sep).join('/');
+        if (rel === 'audit-wiring.test.ts') continue; // this file names the symbol
+        if (rel.startsWith('audit/')) continue;
+        if (/\.audit(\.test)?\.ts$/.test(rel)) continue;
+        if (fs.readFileSync(full, 'utf8').includes('emitAuditEvent')) offenders.push(rel);
+      }
+    };
+    walk(srcRoot);
+    expect(offenders).toEqual([]);
+  });
+});

.claude/skills/add-audit/add/src/audit/config.ts

@@ -0,0 +1,30 @@
+/**
+ * Audit env config (installed by /add-audit) — the two audit vars, read the
+ * same way src/config.ts reads every host setting (readEnvFile from .env,
+ * with process.env taking precedence). Kept audit-owned so installing the
+ * skill never edits core config.ts: the feature's whole footprint in core is
+ * the dispatch composition in dispatch.ts and the resource-barrel import.
+ */
+import { readEnvFile } from '../env.js';
+
+const envConfig = readEnvFile(['AUDIT_ENABLED', 'AUDIT_RETENTION_DAYS']);
+
+/**
+ * Master switch. Off by default — nothing is persisted (and data/audit/ is
+ * never created) until an operator sets AUDIT_ENABLED=true.
+ */
+export const AUDIT_ENABLED = (process.env.AUDIT_ENABLED || envConfig.AUDIT_ENABLED) === 'true';
+
+/**
+ * Day-file retention horizon in days; consulted only when audit is enabled.
+ * Unset or unparseable → 90. An explicit 0 (or negative) = keep forever.
+ */
+function parseRetentionDays(raw: string | undefined): number {
+  if (raw === undefined || raw.trim() === '') return 90;
+  const n = Number.parseInt(raw, 10);
+  return Number.isNaN(n) ? 90 : n;
+}
+
+export const AUDIT_RETENTION_DAYS = parseRetentionDays(
+  process.env.AUDIT_RETENTION_DAYS || envConfig.AUDIT_RETENTION_DAYS,
+);

.claude/skills/add-audit/add/src/audit/emit.ts

@@ -0,0 +1,43 @@
+/**
+ * The single emit seam. Adapters import emitAuditEvent from here directly
+ * (it is deliberately not re-exported by the barrel): only src/audit/ and
+ * `*.audit.ts` adapter files may call it — grep holds the invariant.
+ *
+ * The opt-in check lives here, so the whole feature switches at one point.
+ * Fail-open + loud: a failed append (or a throwing input thunk) is
+ * log.error'd and the audited action proceeds.
+ */
+import { randomUUID } from 'crypto';
+
+import { log } from '../log.js';
+import { AUDIT_ENABLED } from './config.js';
+import { notifyAuditHooks } from './hooks.js';
+import { redactDetails } from './redact.js';
+import { appendAuditLine } from './store.js';
+import type { AuditEvent, AuditEventInput } from './types.js';
+
+export function emitAuditEvent(input: AuditEventInput | (() => AuditEventInput)): void {
+  if (!AUDIT_ENABLED) return; // The one opt-in check — the whole feature switches here.
+  try {
+    if (typeof input === 'function') input = input();
+    const event: AuditEvent = {
+      event_id: randomUUID(),
+      time: new Date().toISOString(),
+      schema_version: 1,
+      actor: { ...input.actor, email: null, user_id: null, group_ids: null },
+      origin: input.origin,
+      action: input.action,
+      resources: input.resources,
+      outcome: input.outcome,
+      correlation_id: input.correlationId ?? null,
+      details: redactDetails(input.details ?? {}),
+    };
+    const line = JSON.stringify(event);
+    appendAuditLine(line);
+    notifyAuditHooks(event, line);
+    // eslint-disable-next-line no-catch-all/no-catch-all -- fail-open is the posture: an audit failure must never take down the audited action
+  } catch (err) {
+    const action = typeof input === 'function' ? undefined : input.action;
+    log.error('Audit append failed — action proceeding (fail-open)', { action, err });
+  }
+}

.claude/skills/add-audit/add/src/audit/hooks.test.ts

@@ -0,0 +1,203 @@
+/**
+ * Post-write hook contract: hooks observe the LOG (fire only after a
+ * successful append, exported ⊆ written), failures are isolated everywhere,
+ * and the lifecycle (init/maintain/shutdown) behaves.
+ */
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+
+const state = vi.hoisted(() => ({ enabled: true, appendThrows: false, appended: [] as string[] }));
+
+vi.mock('../config.js', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('../config.js')>();
+  return {
+    ...actual,
+    DATA_DIR: '/tmp/nanoclaw-test-hooks-unused',
+  };
+});
+
+vi.mock('./config.js', () => ({
+  get AUDIT_ENABLED() {
+    return state.enabled;
+  },
+  AUDIT_RETENTION_DAYS: 90,
+}));
+
+vi.mock('./store.js', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('./store.js')>();
+  return {
+    ...actual,
+    appendAuditLine: (line: string) => {
+      if (state.appendThrows) throw new Error('disk full');
+      state.appended.push(line);
+    },
+  };
+});
+
+vi.mock('../log.js', () => ({
+  log: { debug: vi.fn(), info: vi.fn(), warn: vi.fn(), error: vi.fn(), fatal: vi.fn() },
+}));
+
+let hooks: typeof import('./hooks.js');
+let emit: typeof import('./emit.js');
+let log: (typeof import('../log.js'))['log'];
+
+beforeEach(async () => {
+  state.enabled = true;
+  state.appendThrows = false;
+  state.appended.length = 0;
+  vi.resetModules(); // fresh hook registry per test
+  hooks = await import('./hooks.js');
+  emit = await import('./emit.js');
+  log = (await import('../log.js')).log;
+  vi.clearAllMocks();
+});
+
+afterEach(() => {
+  vi.clearAllMocks();
+});
+
+const EVENT_INPUT = {
+  actor: { type: 'human' as const, id: 'host:test' },
+  origin: { transport: 'socket' as const },
+  action: 'groups.list',
+  resources: [{ type: 'agent_group' }],
+  outcome: 'success' as const,
+  details: { limit: 5 },
+};
+
+describe('post-write notification', () => {
+  it('calls a registered hook with the parsed event and the exact stored line', () => {
+    const seen: Array<{ event: import('./types.js').AuditEvent; line: string }> = [];
+    hooks.registerAuditHook({ name: 'demo', onEvent: (event, line) => seen.push({ event, line }) });
+
+    emit.emitAuditEvent(EVENT_INPUT);
+
+    expect(state.appended).toHaveLength(1);
+    expect(seen).toHaveLength(1);
+    expect(seen[0].line).toBe(state.appended[0]);
+    expect(JSON.parse(seen[0].line)).toEqual(seen[0].event);
+    expect(seen[0].event.action).toBe('groups.list');
+  });
+
+  it('does NOT call hooks when the local append fails — exported ⊆ written', () => {
+    const onEvent = vi.fn();
+    hooks.registerAuditHook({ name: 'demo', onEvent });
+    state.appendThrows = true;
+
+    expect(() => emit.emitAuditEvent(EVENT_INPUT)).not.toThrow(); // action still proceeds
+    expect(onEvent).not.toHaveBeenCalled();
+    expect(log.error).toHaveBeenCalledWith(expect.stringContaining('Audit append failed'), expect.anything());
+  });
+
+  it('does NOT call hooks when audit is disabled', () => {
+    const onEvent = vi.fn();
+    hooks.registerAuditHook({ name: 'demo', onEvent });
+    state.enabled = false;
+
+    emit.emitAuditEvent(EVENT_INPUT);
+
+    expect(state.appended).toHaveLength(0);
+    expect(onEvent).not.toHaveBeenCalled();
+  });
+
+  it('isolates a throwing hook: the write survives, later hooks still run, the action proceeds', () => {
+    const second = vi.fn();
+    hooks.registerAuditHook({
+      name: 'broken',
+      onEvent: () => {
+        throw new Error('exporter exploded');
+      },
+    });
+    hooks.registerAuditHook({ name: 'healthy', onEvent: second });
+
+    expect(() => emit.emitAuditEvent(EVENT_INPUT)).not.toThrow();
+
+    expect(state.appended).toHaveLength(1); // the log has the event regardless
+    expect(second).toHaveBeenCalledTimes(1);
+    expect(log.error).toHaveBeenCalledWith(
+      expect.stringContaining('Audit hook threw'),
+      expect.objectContaining({ hook: 'broken', action: 'groups.list' }),
+    );
+  });
+});
+
+describe('lifecycle', () => {
+  it('initAuditHooks surfaces a failing init as a fatal error naming the hook', () => {
+    hooks.registerAuditHook({ name: 'ok', onEvent: () => {}, init: vi.fn() });
+    hooks.registerAuditHook({
+      name: 'bad-boot',
+      onEvent: () => {},
+      init: () => {
+        throw new Error('no route to collector');
+      },
+    });
+
+    expect(() => hooks.initAuditHooks()).toThrow(/audit hook "bad-boot" failed to initialize.*no route/);
+  });
+
+  it('maintainAuditHooks calls every maintain and isolates throws', () => {
+    const m1 = vi.fn(() => {
+      throw new Error('flush failed');
+    });
+    const m2 = vi.fn();
+    hooks.registerAuditHook({ name: 'a', onEvent: () => {}, maintain: m1 });
+    hooks.registerAuditHook({ name: 'b', onEvent: () => {}, maintain: m2 });
+    hooks.registerAuditHook({ name: 'c', onEvent: () => {} }); // no maintain — fine
+
+    expect(() => hooks.maintainAuditHooks()).not.toThrow();
+    expect(m1).toHaveBeenCalledTimes(1);
+    expect(m2).toHaveBeenCalledTimes(1);
+    expect(log.error).toHaveBeenCalledWith(
+      expect.stringContaining('maintenance failed'),
+      expect.objectContaining({ hook: 'a' }),
+    );
+  });
+
+  it('shutdownAuditHooks awaits async shutdowns and isolates throws', async () => {
+    const order: string[] = [];
+    hooks.registerAuditHook({
+      name: 'a',
+      onEvent: () => {},
+      shutdown: async () => {
+        await Promise.resolve();
+        order.push('a');
+      },
+    });
+    hooks.registerAuditHook({
+      name: 'b',
+      onEvent: () => {},
+      shutdown: () => {
+        throw new Error('handle already closed');
+      },
+    });
+    hooks.registerAuditHook({
+      name: 'c',
+      onEvent: () => {},
+      shutdown: () => {
+        order.push('c');
+      },
+    });
+
+    await hooks.shutdownAuditHooks();
+
+    expect(order).toEqual(['a', 'c']);
+    expect(log.error).toHaveBeenCalledWith(
+      expect.stringContaining('shutdown failed'),
+      expect.objectContaining({ hook: 'b' }),
+    );
+  });
+
+  it('maintainAudit skips hook maintenance when audit is disabled', async () => {
+    const init = await import('./init.js');
+    const maintain = vi.fn();
+    hooks.registerAuditHook({ name: 'a', onEvent: () => {}, maintain });
+
+    state.enabled = false;
+    init.maintainAudit();
+    expect(maintain).not.toHaveBeenCalled();
+
+    state.enabled = true;
+    init.maintainAudit();
+    expect(maintain).toHaveBeenCalledTimes(1);
+  });
+});

.claude/skills/add-audit/add/src/audit/hooks.ts

@@ -0,0 +1,92 @@
+/**
+ * Post-write audit hooks — the in-process extension seam.
+ *
+ * A hook observes the audit LOG, not the event stream: `onEvent` fires only
+ * after an event has been durably appended to the local day-file, so anything
+ * a hook exports is guaranteed to exist in the source of truth
+ * (exported ⊆ written). If the local append fails, hooks are not called —
+ * and a hook that misses events (crash, restart) catches up by reading the
+ * day-files, which is the at-least-once story.
+ *
+ * Registration follows the tree's observer idiom (registerApprovalResolvedHandler,
+ * registerResponseHandler, …): an in-tree or skill-installed module calls
+ * `registerAuditHook(...)` at import time — no core edits, and credentials or
+ * transport for an external system live in that module, never here.
+ */
+import { log } from '../log.js';
+import type { AuditEvent } from './types.js';
+
+export interface AuditHook {
+  /** Short identifier used in logs and lifecycle errors. */
+  name: string;
+  /**
+   * Called after a successful local append. `line` is the exact stored bytes
+   * (one NDJSON line, no trailing newline); `event` is the parsed record.
+   * MUST be fast and non-blocking — this runs on the audited action's call
+   * path. A real exporter buffers here and does its IO from `maintain`/its own
+   * timers. Throwing is tolerated: isolated and logged, never propagated.
+   */
+  onEvent(event: AuditEvent, line: string): void;
+  /** Boot hook, called only when audit is enabled. Throw = host refuses to start. */
+  init?(): void;
+  /** Periodic maintenance — called from the audit maintenance timer (enabled boxes only). */
+  maintain?(): void;
+  /** Graceful-shutdown hook (flush buffers, close handles). */
+  shutdown?(): void | Promise<void>;
+}
+
+const hooks: AuditHook[] = [];
+
+export function registerAuditHook(hook: AuditHook): void {
+  hooks.push(hook);
+}
+
+/** Fan out one written event to every hook, isolating failures per hook. */
+export function notifyAuditHooks(event: AuditEvent, line: string): void {
+  for (const hook of hooks) {
+    try {
+      hook.onEvent(event, line);
+      // eslint-disable-next-line no-catch-all/no-catch-all -- isolation is the contract: one bad hook must not affect the log, other hooks, or the audited action
+    } catch (err) {
+      log.error('Audit hook threw — event is safely in the log', { hook: hook.name, action: event.action, err });
+    }
+  }
+}
+
+/** Boot lifecycle. A hook that can't start is a silent-export-gap risk — fatal. */
+export function initAuditHooks(): void {
+  for (const hook of hooks) {
+    try {
+      hook.init?.();
+    } catch (err) {
+      throw new Error(
+        `audit hook "${hook.name}" failed to initialize: ${err instanceof Error ? err.message : String(err)}`,
+        { cause: err },
+      );
+    }
+  }
+}
+
+/** Periodic lifecycle — maintenance, isolated per hook. */
+export function maintainAuditHooks(): void {
+  for (const hook of hooks) {
+    try {
+      hook.maintain?.();
+      // eslint-disable-next-line no-catch-all/no-catch-all -- one hook's maintenance failure must not stop the others (or the timer)
+    } catch (err) {
+      log.error('Audit hook maintenance failed', { hook: hook.name, err });
+    }
+  }
+}
+
+/** Shutdown lifecycle — awaited by the host's graceful shutdown. */
+export async function shutdownAuditHooks(): Promise<void> {
+  for (const hook of hooks) {
+    try {
+      await hook.shutdown?.();
+      // eslint-disable-next-line no-catch-all/no-catch-all -- shutdown must drain every hook even when one throws
+    } catch (err) {
+      log.error('Audit hook shutdown failed', { hook: hook.name, err });
+    }
+  }
+}

.claude/skills/add-audit/add/src/audit/index.ts

@@ -0,0 +1,18 @@
+/**
+ * Opt-in local audit log — installed by /add-audit.
+ *
+ * This directory is a domain-free leaf: the event schema, the emit seam, the
+ * store, the reader, post-write hooks, and shared vocabulary. What gets
+ * audited — and how each domain describes itself — lives in the domain-owned
+ * `*.audit.ts` adapter files next to the code they observe. Business logic
+ * contains zero audit calls.
+ *
+ * emitAuditEvent is deliberately NOT re-exported here: adapters import it
+ * from './emit.js' directly, and only src/audit/ + `*.audit.ts` may call it.
+ */
+export * from './types.js';
+export { redactDetails } from './redact.js';
+export { AUDIT_DIR } from './store.js';
+export { AUDIT_ENABLED, AUDIT_RETENTION_DAYS } from './config.js';
+export { initAuditLog, maintainAudit } from './init.js';
+export { type AuditHook, registerAuditHook } from './hooks.js';

.claude/skills/add-audit/add/src/audit/init.ts

@@ -0,0 +1,54 @@
+/**
+ * Boot-time audit wiring, self-contained (installed by /add-audit).
+ *
+ * initAuditLog() runs at module load of the CLI audit adapter
+ * (cli/dispatch.audit.ts) — i.e. during the host's barrel phase, before the
+ * CLI server or any delivery poll accepts work — because dispatch.ts, which
+ * both transports import, composes withAudit at module scope. When enabled:
+ * assert data/audit/ is writable (refusing to start beats running with a
+ * silent audit gap), run the boot prune, start the registered post-write
+ * hooks, and arm the maintenance timer.
+ *
+ * The timer owns the daily cadence in-module (checked hourly; the prune
+ * itself fires at most once per UTC day) so installing the skill touches
+ * dispatch.ts and the resource barrel only — no host-sweep edit. It is
+ * unref'd: it never keeps the process alive.
+ */
+import { log } from '../log.js';
+import { onShutdown } from '../response-registry.js';
+import { AUDIT_ENABLED, AUDIT_RETENTION_DAYS } from './config.js';
+import { initAuditHooks, maintainAuditHooks, shutdownAuditHooks } from './hooks.js';
+import { assertAuditWritable, AUDIT_DIR, markPrunedToday, pruneAuditLog, pruneAuditLogIfDue } from './store.js';
+
+const MAINTENANCE_INTERVAL_MS = 60 * 60 * 1000;
+
+let initialized = false;
+
+export function initAuditLog(): void {
+  if (!AUDIT_ENABLED || initialized) return;
+  initialized = true;
+  try {
+    assertAuditWritable();
+  } catch (err) {
+    throw new Error(
+      `AUDIT_ENABLED=true but the audit directory is not writable: ${AUDIT_DIR} (${err instanceof Error ? err.message : String(err)})`,
+      { cause: err },
+    );
+  }
+  pruneAuditLog();
+  markPrunedToday();
+  initAuditHooks(); // throw → boot fails, same posture as the writability assert
+  onShutdown(() => shutdownAuditHooks());
+  setInterval(maintainAudit, MAINTENANCE_INTERVAL_MS).unref();
+  log.info('Audit log enabled', { dir: AUDIT_DIR, retentionDays: AUDIT_RETENTION_DAYS });
+}
+
+/**
+ * One maintenance tick: retention prune (throttled to once per UTC day
+ * internally) plus every hook's periodic maintenance. No-op when audit is
+ * disabled. Exported so a future scheduler can drive it too.
+ */
+export function maintainAudit(): void {
+  pruneAuditLogIfDue();
+  if (AUDIT_ENABLED) maintainAuditHooks();
+}

.claude/skills/add-audit/add/src/audit/reader.test.ts

@@ -0,0 +1,153 @@
+import fs from 'fs';
+import os from 'os';
+import path from 'path';
+
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+
+const state = vi.hoisted(() => ({ dataDir: '', enabled: true }));
+
+vi.mock('../config.js', () => ({
+  get DATA_DIR() {
+    return state.dataDir;
+  },
+}));
+
+vi.mock('./config.js', () => ({
+  get AUDIT_ENABLED() {
+    return state.enabled;
+  },
+  AUDIT_RETENTION_DAYS: 90,
+}));
+
+vi.mock('../log.js', () => ({
+  log: { debug: vi.fn(), info: vi.fn(), warn: vi.fn(), error: vi.fn(), fatal: vi.fn() },
+}));
+
+let reader: typeof import('./reader.js');
+let store: typeof import('./store.js');
+
+beforeEach(async () => {
+  state.dataDir = fs.mkdtempSync(path.join(os.tmpdir(), 'audit-reader-'));
+  state.enabled = true;
+  vi.resetModules();
+  store = await import('./store.js');
+  reader = await import('./reader.js');
+});
+
+afterEach(() => {
+  fs.rmSync(state.dataDir, { recursive: true, force: true });
+});
+
+function dayString(daysAgo: number): string {
+  return store.utcDay(new Date(Date.now() - daysAgo * 24 * 60 * 60 * 1000));
+}
+
+function isoAt(daysAgo: number, tag: number): string {
+  const base = new Date(Date.now() - daysAgo * 24 * 60 * 60 * 1000);
+  return `${base.toISOString().slice(0, 10)}T10:00:0${tag}.000Z`;
+}
+
+let seq = 0;
+function seedEvent(daysAgo: number, overrides: Record<string, unknown> = {}): Record<string, unknown> {
+  seq += 1;
+  const event = {
+    event_id: `e-${seq}`,
+    time: isoAt(daysAgo, seq % 10),
+    schema_version: 1,
+    actor: { type: 'human', id: 'host:moshe', email: null, user_id: null, group_ids: null },
+    origin: { transport: 'socket' },
+    action: 'groups.list',
+    resources: [{ type: 'agent_group', id: 'ag-1' }],
+    outcome: 'success',
+    correlation_id: null,
+    details: {},
+    ...overrides,
+  };
+  const dir = path.join(state.dataDir, 'audit');
+  fs.mkdirSync(dir, { recursive: true });
+  fs.appendFileSync(path.join(dir, `${dayString(daysAgo)}.ndjson`), JSON.stringify(event) + '\n');
+  return event;
+}
+
+describe('listAuditEvents', () => {
+  it('throws the disabled error rather than returning an empty list', () => {
+    state.enabled = false;
+    expect(() => reader.listAuditEvents({})).toThrow('audit log is disabled — set AUDIT_ENABLED=true');
+  });
+
+  it('returns flat rows newest-first across day-files, honoring --limit', () => {
+    seedEvent(2, { event_id: 'old' });
+    seedEvent(1, { event_id: 'mid-a' });
+    seedEvent(1, { event_id: 'mid-b' });
+    seedEvent(0, { event_id: 'new' });
+
+    const rows = reader.listAuditEvents({}) as Array<Record<string, unknown>>;
+    expect(rows.map((r) => r.event_id)).toEqual(['new', 'mid-b', 'mid-a', 'old']);
+    expect(rows[0]).toMatchObject({ actor: 'host:moshe', action: 'groups.list', outcome: 'success' });
+    expect(rows[0].resources).toBe('agent_group:ag-1');
+
+    const limited = reader.listAuditEvents({ limit: 2 }) as Array<Record<string, unknown>>;
+    expect(limited.map((r) => r.event_id)).toEqual(['new', 'mid-b']);
+  });
+
+  it('filters by actor, outcome, correlation, and resource (id or type)', () => {
+    seedEvent(0, { event_id: 'a', actor: { type: 'agent', id: 'ag-1' }, outcome: 'denied' });
+    seedEvent(0, { event_id: 'b', correlation_id: 'appr-9', resources: [{ type: 'approval', id: 'appr-9' }] });
+    seedEvent(0, { event_id: 'c', resources: [{ type: 'user', id: 'slack:U1' }] });
+
+    expect((reader.listAuditEvents({ actor: 'ag-1' }) as unknown[]).length).toBe(1);
+    expect((reader.listAuditEvents({ outcome: 'denied' }) as unknown[]).length).toBe(1);
+    expect((reader.listAuditEvents({ correlation: 'appr-9' }) as unknown[]).length).toBe(1);
+    expect((reader.listAuditEvents({ resource: 'slack:U1' }) as unknown[]).length).toBe(1);
+    expect((reader.listAuditEvents({ resource: 'approval' }) as unknown[]).length).toBe(1);
+  });
+
+  it('matches actions exactly or by dotted prefix', () => {
+    seedEvent(0, { event_id: 'cfg', action: 'groups.config.add-mcp-server' });
+    seedEvent(0, { event_id: 'list', action: 'groups.list' });
+    seedEvent(0, { event_id: 'other', action: 'sessions.list' });
+
+    expect((reader.listAuditEvents({ action: 'groups' }) as unknown[]).length).toBe(2);
+    expect((reader.listAuditEvents({ action: 'groups.config' }) as unknown[]).length).toBe(1);
+    expect((reader.listAuditEvents({ action: 'groups.list' }) as unknown[]).length).toBe(1);
+    // Prefix means dotted segments, not substrings.
+    expect((reader.listAuditEvents({ action: 'group' }) as unknown[]).length).toBe(0);
+  });
+
+  it('applies --since/--until with relative and ISO forms', () => {
+    seedEvent(5, { event_id: 'old' });
+    seedEvent(0, { event_id: 'recent' });
+
+    const relative = reader.listAuditEvents({ since: '2d' }) as Array<Record<string, unknown>>;
+    expect(relative.map((r) => r.event_id)).toEqual(['recent']);
+
+    const iso = reader.listAuditEvents({ until: dayString(2) }) as Array<Record<string, unknown>>;
+    expect(iso.map((r) => r.event_id)).toEqual(['old']);
+
+    expect(() => reader.listAuditEvents({ since: 'yesterdayish' })).toThrow('invalid --since');
+  });
+
+  it('--format ndjson returns the stored lines verbatim', () => {
+    const seeded = seedEvent(0, { event_id: 'x1' });
+    const out = reader.listAuditEvents({ format: 'ndjson' });
+    expect(typeof out).toBe('string');
+    expect(JSON.parse(out as string)).toEqual(seeded);
+    expect(() => reader.listAuditEvents({ format: 'csv' })).toThrow('invalid --format');
+  });
+
+  it('skips malformed stored lines and still returns the rest', () => {
+    seedEvent(0, { event_id: 'good' });
+    fs.appendFileSync(path.join(state.dataDir, 'audit', `${dayString(0)}.ndjson`), 'not-json\n');
+
+    const rows = reader.listAuditEvents({}) as Array<Record<string, unknown>>;
+    expect(rows.map((r) => r.event_id)).toEqual(['good']);
+  });
+
+  it('rejects an unknown --outcome value', () => {
+    expect(() => reader.listAuditEvents({ outcome: 'meh' })).toThrow('invalid --outcome');
+  });
+
+  it('returns empty when nothing has been recorded yet', () => {
+    expect(reader.listAuditEvents({})).toEqual([]);
+  });
+});

.claude/skills/add-audit/add/src/audit/reader.ts

@@ -0,0 +1,160 @@
+/**
+ * Read-back for `ncl audit list` — a newest-first stream-scan over the
+ * day-files. No index: fine at v1 volume, and adding one later doesn't change
+ * the store. NDJSON export returns the stored lines verbatim.
+ */
+import fs from 'fs';
+import path from 'path';
+
+import { log } from '../log.js';
+import { AUDIT_ENABLED } from './config.js';
+import { AUDIT_DIR, utcDay } from './store.js';
+import type { AuditEvent, AuditOutcome } from './types.js';
+
+const OUTCOMES: ReadonlySet<string> = new Set(['success', 'failure', 'denied', 'pending', 'approved', 'rejected']);
+const DAY_FILE_RE = /^(\d{4}-\d{2}-\d{2})\.ndjson$/;
+const DEFAULT_LIMIT = 100;
+
+export interface AuditQuery {
+  actor?: string;
+  /** Exact action or dotted prefix (`groups` matches `groups.config.update`). */
+  action?: string;
+  /** Matches any resources[] entry by id or by type. */
+  resource?: string;
+  outcome?: AuditOutcome;
+  sinceMs?: number;
+  untilMs?: number;
+  correlation?: string;
+  limit: number;
+}
+
+/** `7d` / `24h` / `30m` relative to now, or an ISO date/datetime (UTC). */
+export function parseTimeFlag(value: string, flag: string): number {
+  const rel = /^(\d+)([dhm])$/.exec(value);
+  if (rel) {
+    const n = Number(rel[1]);
+    const unitMs = rel[2] === 'd' ? 86_400_000 : rel[2] === 'h' ? 3_600_000 : 60_000;
+    return Date.now() - n * unitMs;
+  }
+  const abs = Date.parse(value);
+  if (!Number.isNaN(abs)) return abs;
+  throw new Error(`invalid ${flag} value "${value}" — use e.g. 7d, 24h, 30m, or an ISO date`);
+}
+
+/** Newest first across files and within each file, up to q.limit. */
+export function queryAuditEvents(q: AuditQuery): { events: AuditEvent[]; lines: string[] } {
+  const events: AuditEvent[] = [];
+  const lines: string[] = [];
+  let malformed = 0;
+
+  for (const { day, file } of dayFilesNewestFirst()) {
+    if (events.length >= q.limit) break;
+    // Whole-day skip: a file can't match a window its day lies outside.
+    if (q.sinceMs !== undefined && day < utcDay(new Date(q.sinceMs))) continue;
+    if (q.untilMs !== undefined && day > utcDay(new Date(q.untilMs))) continue;
+
+    let content: string;
+    try {
+      content = fs.readFileSync(file, 'utf8');
+      // eslint-disable-next-line no-catch-all/no-catch-all -- a torn/pruned day-file must not fail the whole query
+    } catch (err) {
+      log.warn('Audit reader failed to read day-file', { file, err });
+      continue;
+    }
+    const fileLines = content.split('\n').filter((l) => l.trim() !== '');
+    // Lines within a file are chronological — walk backwards for newest-first.
+    for (let i = fileLines.length - 1; i >= 0 && events.length < q.limit; i--) {
+      let event: AuditEvent;
+      try {
+        event = JSON.parse(fileLines[i]) as AuditEvent;
+        // eslint-disable-next-line no-catch-all/no-catch-all -- malformed stored lines are skipped (counted + warned below)
+      } catch {
+        malformed++;
+        continue;
+      }
+      if (!matches(event, q)) continue;
+      events.push(event);
+      lines.push(fileLines[i]);
+    }
+  }
+
+  if (malformed > 0) {
+    log.warn('Audit reader skipped malformed lines', { malformed });
+  }
+  return { events, lines };
+}
+
+function dayFilesNewestFirst(): Array<{ day: string; file: string }> {
+  let entries: string[];
+  try {
+    entries = fs.readdirSync(AUDIT_DIR);
+    // eslint-disable-next-line no-catch-all/no-catch-all -- no audit dir yet means no events, not an error
+  } catch {
+    return [];
+  }
+  return entries
+    .map((e) => DAY_FILE_RE.exec(e))
+    .filter((m): m is RegExpExecArray => m !== null)
+    .map((m) => ({ day: m[1], file: path.join(AUDIT_DIR, m[0]) }))
+    .sort((a, b) => (a.day < b.day ? 1 : -1));
+}
+
+function matches(event: AuditEvent, q: AuditQuery): boolean {
+  if (q.actor !== undefined && event.actor?.id !== q.actor) return false;
+  if (q.action !== undefined && event.action !== q.action && !event.action?.startsWith(q.action + '.')) return false;
+  if (q.outcome !== undefined && event.outcome !== q.outcome) return false;
+  if (q.correlation !== undefined && event.correlation_id !== q.correlation) return false;
+  if (q.resource !== undefined) {
+    const hit = (event.resources ?? []).some((r) => r.id === q.resource || r.type === q.resource);
+    if (!hit) return false;
+  }
+  const t = Date.parse(event.time ?? '');
+  if (q.sinceMs !== undefined && !(t >= q.sinceMs)) return false;
+  if (q.untilMs !== undefined && !(t <= q.untilMs)) return false;
+  return true;
+}
+
+/**
+ * `ncl audit list` handler. Disabled → an explicit error: an empty list would
+ * read as "no actions happened", which is a different truth than "not
+ * recording". `--format ndjson` returns the stored lines verbatim (the human
+ * formatter passes strings through); default returns flat rows for the table.
+ */
+export function listAuditEvents(args: Record<string, unknown>): string | Array<Record<string, unknown>> {
+  if (!AUDIT_ENABLED) {
+    throw new Error('audit log is disabled — set AUDIT_ENABLED=true');
+  }
+
+  const format = args.format !== undefined ? String(args.format) : '';
+  if (format && format !== 'ndjson') {
+    throw new Error(`invalid --format "${format}" — only "ndjson" is supported`);
+  }
+  const outcome = args.outcome !== undefined ? String(args.outcome) : undefined;
+  if (outcome !== undefined && !OUTCOMES.has(outcome)) {
+    throw new Error(`invalid --outcome "${outcome}" — one of: ${[...OUTCOMES].join(', ')}`);
+  }
+
+  const q: AuditQuery = {
+    actor: args.actor !== undefined ? String(args.actor) : undefined,
+    action: args.action !== undefined ? String(args.action) : undefined,
+    resource: args.resource !== undefined ? String(args.resource) : undefined,
+    outcome: outcome as AuditOutcome | undefined,
+    correlation: args.correlation !== undefined ? String(args.correlation) : undefined,
+    sinceMs: args.since !== undefined ? parseTimeFlag(String(args.since), '--since') : undefined,
+    untilMs: args.until !== undefined ? parseTimeFlag(String(args.until), '--until') : undefined,
+    limit: args.limit !== undefined ? Math.max(1, Number(args.limit) || DEFAULT_LIMIT) : DEFAULT_LIMIT,
+  };
+
+  const { events, lines } = queryAuditEvents(q);
+  if (format === 'ndjson') return lines.join('\n');
+
+  return events.map((e) => ({
+    time: e.time,
+    actor: e.actor?.id ?? '',
+    action: e.action,
+    resources: (e.resources ?? []).map((r) => (r.id ? `${r.type}:${r.id}` : r.type)).join(' '),
+    outcome: e.outcome,
+    correlation: e.correlation_id ?? '',
+    event_id: e.event_id,
+  }));
+}

.claude/skills/add-audit/add/src/audit/redact.test.ts

@@ -0,0 +1,69 @@
+import { describe, expect, it } from 'vitest';
+
+import { redactDetails } from './redact.js';
+
+describe('redactDetails', () => {
+  it('masks sensitive keys at any depth, case-insensitively', () => {
+    const out = redactDetails({
+      name: 'notion',
+      env: { NOTION_TOKEN: 'secret-value', SAFE_VALUE: 'ok' },
+      nested: { Authorization: 'Bearer abc', list: [{ 'api-key': 'k' }, { plain: 'p' }] },
+      password: 'hunter2',
+    });
+    expect(out).toEqual({
+      name: 'notion',
+      env: { NOTION_TOKEN: '[REDACTED]', SAFE_VALUE: 'ok' },
+      nested: { Authorization: '[REDACTED]', list: [{ 'api-key': '[REDACTED]' }, { plain: 'p' }] },
+      password: '[REDACTED]',
+    });
+  });
+
+  it('matches the documented key pattern (token|secret|key|password|credential|auth|bearer)', () => {
+    const out = redactDetails({
+      access_token: 'x',
+      clientSecret: 'x',
+      ssh_key: 'x',
+      credentials: 'x',
+      oauth_flow: 'x',
+      bearerValue: 'x',
+      username: 'moshe',
+    });
+    expect(Object.entries(out).filter(([, v]) => v === '[REDACTED]')).toHaveLength(6);
+    expect(out.username).toBe('moshe');
+  });
+
+  it('never recurses into a masked key — the whole value is replaced', () => {
+    const out = redactDetails({ auth: { inner: 'visible?' } });
+    expect(out.auth).toBe('[REDACTED]');
+  });
+
+  it('truncates strings over 2 KB post-redaction', () => {
+    const long = 'a'.repeat(5000);
+    const out = redactDetails({ blob: long, short: 'b' });
+    expect(out.blob).toBe('a'.repeat(2048) + '…[truncated]');
+    expect(out.short).toBe('b');
+  });
+
+  it('passes non-string scalars through untouched', () => {
+    const out = redactDetails({ n: 42, b: true, z: null, u: undefined });
+    expect(out).toEqual({ n: 42, b: true, z: null, u: undefined });
+  });
+
+  it('caps depth (cycle guard) without throwing', () => {
+    const cyclic: Record<string, unknown> = { a: 1 };
+    cyclic.self = cyclic;
+    const out = redactDetails(cyclic);
+    let cursor: unknown = out;
+    for (let i = 0; i < 20 && typeof cursor === 'object' && cursor !== null; i++) {
+      cursor = (cursor as Record<string, unknown>).self;
+    }
+    expect(cursor).toBe('[MAX_DEPTH]');
+  });
+
+  it('does not mutate the input', () => {
+    const input = { token: 'x', nested: { list: ['a'.repeat(3000)] } };
+    const snapshot = JSON.parse(JSON.stringify(input));
+    redactDetails(input);
+    expect(input).toEqual(snapshot);
+  });
+});

.claude/skills/add-audit/add/src/audit/redact.ts

@@ -0,0 +1,32 @@
+/**
+ * The one redactor, applied to every event's `details` at the emit seam —
+ * guarding current and future surfaces by default. Key-pattern mask plus a
+ * per-value size cap; message bodies are never passed in (emit sites record
+ * shape only — the audit log is a governance record, not a chat archive).
+ */
+
+const SENSITIVE_KEY = /(token|secret|key|password|credential|auth|bearer)/i;
+const MAX_VALUE_CHARS = 2048;
+const MAX_DEPTH = 8;
+
+export function redactDetails(details: Record<string, unknown>): Record<string, unknown> {
+  return redactObject(details, 0);
+}
+
+function redactObject(obj: Record<string, unknown>, depth: number): Record<string, unknown> {
+  const out: Record<string, unknown> = {};
+  for (const [k, v] of Object.entries(obj)) {
+    out[k] = SENSITIVE_KEY.test(k) ? '[REDACTED]' : redactValue(v, depth + 1);
+  }
+  return out;
+}
+
+function redactValue(value: unknown, depth: number): unknown {
+  if (typeof value === 'string') {
+    return value.length > MAX_VALUE_CHARS ? `${value.slice(0, MAX_VALUE_CHARS)}…[truncated]` : value;
+  }
+  if (value === null || typeof value !== 'object') return value;
+  if (depth > MAX_DEPTH) return '[MAX_DEPTH]';
+  if (Array.isArray(value)) return value.map((v) => redactValue(v, depth + 1));
+  return redactObject(value as Record<string, unknown>, depth);
+}

.claude/skills/add-audit/add/src/audit/store.test.ts

@@ -0,0 +1,187 @@
+import fs from 'fs';
+import os from 'os';
+import path from 'path';
+
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+
+// Core config (DATA_DIR) and the audit config (toggles) are mocked so
+// store/emit resolve a per-test temp DATA_DIR and audit switches. Getters
+// keep the values live across vi.resetModules().
+const state = vi.hoisted(() => ({ dataDir: '', enabled: true, retention: 90 }));
+
+vi.mock('../config.js', () => ({
+  get DATA_DIR() {
+    return state.dataDir;
+  },
+}));
+
+vi.mock('./config.js', () => ({
+  get AUDIT_ENABLED() {
+    return state.enabled;
+  },
+  get AUDIT_RETENTION_DAYS() {
+    return state.retention;
+  },
+}));
+
+vi.mock('../log.js', () => ({
+  log: { debug: vi.fn(), info: vi.fn(), warn: vi.fn(), error: vi.fn(), fatal: vi.fn() },
+}));
+
+let store: typeof import('./store.js');
+let emit: typeof import('./emit.js');
+let log: (typeof import('../log.js'))['log'];
+
+beforeEach(async () => {
+  state.dataDir = fs.mkdtempSync(path.join(os.tmpdir(), 'audit-store-'));
+  state.enabled = true;
+  state.retention = 90;
+  vi.resetModules();
+  store = await import('./store.js');
+  emit = await import('./emit.js');
+  log = (await import('../log.js')).log;
+  vi.clearAllMocks();
+});
+
+afterEach(() => {
+  fs.chmodSync(path.join(state.dataDir), 0o700);
+  const auditDir = path.join(state.dataDir, 'audit');
+  if (fs.existsSync(auditDir)) fs.chmodSync(auditDir, 0o700);
+  fs.rmSync(state.dataDir, { recursive: true, force: true });
+});
+
+function auditDir(): string {
+  return path.join(state.dataDir, 'audit');
+}
+
+function writeDayFile(day: string, lines = 1): void {
+  fs.mkdirSync(auditDir(), { recursive: true });
+  fs.writeFileSync(path.join(auditDir(), `${day}.ndjson`), '{"x":1}\n'.repeat(lines));
+}
+
+function dayString(daysAgo: number): string {
+  return store.utcDay(new Date(Date.now() - daysAgo * 24 * 60 * 60 * 1000));
+}
+
+const EVENT_INPUT = {
+  actor: { type: 'human' as const, id: 'host:test' },
+  origin: { transport: 'socket' as const },
+  action: 'groups.list',
+  resources: [{ type: 'agent_group' }],
+  outcome: 'success' as const,
+};
+
+describe('appendAuditLine', () => {
+  it("appends one line to today's UTC day-file, creating the directory lazily", () => {
+    expect(fs.existsSync(auditDir())).toBe(false);
+    store.appendAuditLine('{"a":1}');
+    store.appendAuditLine('{"b":2}');
+    const file = path.join(auditDir(), `${store.utcDay()}.ndjson`);
+    expect(fs.readFileSync(file, 'utf8')).toBe('{"a":1}\n{"b":2}\n');
+  });
+});
+
+describe('emitAuditEvent', () => {
+  it('writes a schema_version-1 record with envelope fields and null enrichment', () => {
+    emit.emitAuditEvent({ ...EVENT_INPUT, details: { limit: 100 } });
+    const file = path.join(auditDir(), `${store.utcDay()}.ndjson`);
+    const record = JSON.parse(fs.readFileSync(file, 'utf8').trim());
+    expect(record).toMatchObject({
+      schema_version: 1,
+      actor: { type: 'human', id: 'host:test', email: null, user_id: null, group_ids: null },
+      origin: { transport: 'socket' },
+      action: 'groups.list',
+      outcome: 'success',
+      correlation_id: null,
+      details: { limit: 100 },
+    });
+    expect(record.event_id).toMatch(/^[0-9a-f-]{36}$/);
+    expect(new Date(record.time).toISOString()).toBe(record.time);
+  });
+
+  it('redacts details at the emit seam', () => {
+    emit.emitAuditEvent({ ...EVENT_INPUT, details: { env: { API_TOKEN: 'x' } } });
+    const file = path.join(auditDir(), `${store.utcDay()}.ndjson`);
+    expect(fs.readFileSync(file, 'utf8')).toContain('"API_TOKEN":"[REDACTED]"');
+  });
+
+  it('is a no-op when audit is disabled — the directory is never created', () => {
+    state.enabled = false;
+    emit.emitAuditEvent(EVENT_INPUT);
+    expect(fs.existsSync(auditDir())).toBe(false);
+  });
+
+  it('fails open and loud when the append fails', () => {
+    fs.mkdirSync(auditDir(), { recursive: true });
+    fs.chmodSync(auditDir(), 0o500);
+    expect(() => emit.emitAuditEvent(EVENT_INPUT)).not.toThrow();
+    expect(log.error).toHaveBeenCalledWith(
+      expect.stringContaining('Audit append failed'),
+      expect.objectContaining({ action: 'groups.list' }),
+    );
+  });
+});
+
+describe('assertAuditWritable', () => {
+  it('creates the directory and probes it with a zero-byte append', () => {
+    store.assertAuditWritable();
+    expect(fs.existsSync(path.join(auditDir(), `${store.utcDay()}.ndjson`))).toBe(true);
+  });
+
+  it('throws when the directory is not writable', () => {
+    fs.mkdirSync(auditDir(), { recursive: true });
+    fs.chmodSync(auditDir(), 0o500);
+    expect(() => store.assertAuditWritable()).toThrow();
+  });
+});
+
+describe('pruneAuditLog', () => {
+  it('unlinks only day-files strictly older than the horizon', () => {
+    writeDayFile(dayString(100));
+    writeDayFile(dayString(91));
+    writeDayFile(dayString(90));
+    writeDayFile(dayString(1));
+    fs.writeFileSync(path.join(auditDir(), 'not-a-day-file.txt'), 'keep');
+    fs.writeFileSync(path.join(auditDir(), '2026-01-01.ndjson.bak'), 'keep');
+    store.pruneAuditLog(90);
+    const left = fs.readdirSync(auditDir()).sort();
+    expect(left).toEqual(
+      [`${dayString(90)}.ndjson`, `${dayString(1)}.ndjson`, '2026-01-01.ndjson.bak', 'not-a-day-file.txt'].sort(),
+    );
+  });
+
+  it('keeps forever when retention is 0', () => {
+    writeDayFile(dayString(400));
+    store.pruneAuditLog(0);
+    expect(fs.existsSync(path.join(auditDir(), `${dayString(400)}.ndjson`))).toBe(true);
+  });
+
+  it('is a no-op when the audit directory does not exist', () => {
+    expect(() => store.pruneAuditLog(90)).not.toThrow();
+  });
+});
+
+describe('pruneAuditLogIfDue', () => {
+  it('prunes at most once per UTC day', () => {
+    writeDayFile(dayString(100));
+    store.pruneAuditLogIfDue();
+    expect(fs.existsSync(path.join(auditDir(), `${dayString(100)}.ndjson`))).toBe(false);
+    writeDayFile(dayString(100));
+    store.pruneAuditLogIfDue();
+    expect(fs.existsSync(path.join(auditDir(), `${dayString(100)}.ndjson`))).toBe(true);
+  });
+
+  it('does nothing when audit is disabled', () => {
+    state.enabled = false;
+    writeDayFile(dayString(100));
+    store.pruneAuditLogIfDue();
+    expect(fs.existsSync(path.join(auditDir(), `${dayString(100)}.ndjson`))).toBe(true);
+  });
+
+  it('does nothing after markPrunedToday until the day rolls over', () => {
+    store.markPrunedToday();
+    writeDayFile(dayString(100));
+    store.pruneAuditLogIfDue();
+    expect(fs.existsSync(path.join(auditDir(), `${dayString(100)}.ndjson`))).toBe(true);
+  });
+});

.claude/skills/add-audit/add/src/audit/store.ts

@@ -0,0 +1,81 @@
+/**
+ * Day-file store — daily NDJSON files under data/audit/, host process is the
+ * single writer. Append-only is structural: nothing here can update a line,
+ * and retention is unlinking whole files (a literal hard delete). Both
+ * transports converge host-side, so single-writer holds by construction.
+ *
+ * This module throws on fs failure; the fail-open posture (log.error, action
+ * proceeds) lives in emit.ts, and the boot-time strictness (refuse to start)
+ * lives in init.ts.
+ */
+import fs from 'fs';
+import path from 'path';
+
+import { DATA_DIR } from '../config.js';
+import { log } from '../log.js';
+import { AUDIT_ENABLED, AUDIT_RETENTION_DAYS } from './config.js';
+
+export const AUDIT_DIR = path.join(DATA_DIR, 'audit');
+const DAY_FILE_RE = /^(\d{4}-\d{2}-\d{2})\.ndjson$/;
+const DAY_MS = 24 * 60 * 60 * 1000;
+
+export function utcDay(d: Date = new Date()): string {
+  return d.toISOString().slice(0, 10);
+}
+
+export function dayFilePath(day: string): string {
+  return path.join(AUDIT_DIR, `${day}.ndjson`);
+}
+
+/** The single append point. Throws on fs failure — emitAuditEvent catches. */
+export function appendAuditLine(line: string): void {
+  fs.mkdirSync(AUDIT_DIR, { recursive: true });
+  fs.appendFileSync(dayFilePath(utcDay()), line + '\n');
+}
+
+/** Boot-time writability assert — a zero-byte append is a true write probe. */
+export function assertAuditWritable(): void {
+  fs.mkdirSync(AUDIT_DIR, { recursive: true });
+  fs.appendFileSync(dayFilePath(utcDay()), '');
+}
+
+/** Unlink day-files strictly older than (today UTC − retentionDays). 0 or negative = keep forever. */
+export function pruneAuditLog(retentionDays: number = AUDIT_RETENTION_DAYS): void {
+  if (retentionDays <= 0) return;
+  let entries: string[];
+  try {
+    entries = fs.readdirSync(AUDIT_DIR);
+    // eslint-disable-next-line no-catch-all/no-catch-all -- no audit dir yet means nothing to prune, not an error
+  } catch {
+    return;
+  }
+  const horizon = utcDay(new Date(Date.now() - retentionDays * DAY_MS));
+  let pruned = 0;
+  for (const entry of entries) {
+    const m = DAY_FILE_RE.exec(entry);
+    if (!m || m[1] >= horizon) continue;
+    try {
+      fs.unlinkSync(path.join(AUDIT_DIR, entry));
+      pruned++;
+      // eslint-disable-next-line no-catch-all/no-catch-all -- one stuck file must not stop the prune of the others
+    } catch (err) {
+      log.error('Audit prune failed to unlink day-file', { file: entry, err });
+    }
+  }
+  if (pruned > 0) log.info('Audit retention pruned day-files', { pruned, retentionDays });
+}
+
+let lastPruneDay: string | null = null;
+
+/** Retention prune, throttled to once per UTC day — safe to call every tick. */
+export function pruneAuditLogIfDue(): void {
+  if (!AUDIT_ENABLED || AUDIT_RETENTION_DAYS <= 0) return;
+  const today = utcDay();
+  if (lastPruneDay === today) return;
+  lastPruneDay = today;
+  pruneAuditLog();
+}
+
+export function markPrunedToday(): void {
+  lastPruneDay = utcDay();
+}

.claude/skills/add-audit/add/src/audit/types.ts

@@ -0,0 +1,60 @@
+/**
+ * Canonical audit event — one flat, SIEM-shaped record per action.
+ *
+ * Fields are chosen so they project losslessly onto OCSF and Elastic ECS
+ * (the two schemas SIEMs converge on); the store keeps the neutral shape and
+ * a future forwarder pays the translation once, at the edge. `schema_version`
+ * covers evolution — additive changes only within a version.
+ */
+
+export type AuditActorType = 'human' | 'agent' | 'system';
+
+export interface AuditActor {
+  type: AuditActorType;
+  /** `host:<os-user>` | `<channel>:<handle>` | agent group id | `host` (system). */
+  id: string;
+}
+
+export interface AuditOrigin {
+  transport: 'socket' | 'container' | 'channel';
+  session_id?: string;
+  messaging_group_id?: string;
+  channel?: string;
+}
+
+export interface AuditResource {
+  type: string;
+  /** Omitted when only the attempted type is known (e.g. a denied list). */
+  id?: string;
+}
+
+export type AuditOutcome = 'success' | 'failure' | 'denied' | 'pending' | 'approved' | 'rejected';
+
+/** What emit sites provide. Envelope fields are stamped by emitAuditEvent. */
+export interface AuditEventInput {
+  actor: AuditActor;
+  origin: AuditOrigin;
+  /** Dotted namespaced verb, e.g. `groups.config.add-mcp-server`. */
+  action: string;
+  resources: AuditResource[];
+  outcome: AuditOutcome;
+  /** The approval id on gated chains; null/omitted otherwise. */
+  correlationId?: string | null;
+  details?: Record<string, unknown>;
+}
+
+export interface AuditEvent {
+  event_id: string;
+  time: string;
+  schema_version: 1;
+  actor: AuditActor & { email: null; user_id: null; group_ids: null };
+  origin: AuditOrigin;
+  action: string;
+  resources: AuditResource[];
+  outcome: AuditOutcome;
+  correlation_id: string | null;
+  details: Record<string, unknown>;
+}
+
+/** Actor for timer/sweep-driven outcomes (expiries, startup sweeps). */
+export const SYSTEM_ACTOR: AuditActor = { type: 'system', id: 'host' };

.claude/skills/add-audit/add/src/audit/vocab.ts

@@ -0,0 +1,39 @@
+/**
+ * Domain-free event vocabulary — actor and origin constructors shared by the
+ * domain-owned `*.audit.ts` adapters (the CLI adapter today; approval and
+ * channel adapters attach here in later increments). Pure derivation; nothing
+ * here writes the log.
+ *
+ * Leaf rule: this module (like the rest of src/audit/) may depend on node,
+ * config/log, shared types, and the db read layer — never on src/cli/* or
+ * src/modules/*. Domain-specific mapping (CLI resources, approval payloads)
+ * lives in the adapter file of the domain that owns it.
+ */
+import os from 'os';
+
+import { getMessagingGroup } from '../db/messaging-groups.js';
+import type { AuditOrigin } from './types.js';
+
+/**
+ * Host callers stamp `host:<install user>` daemon-side: the ncl socket is
+ * 0600 and owned by the install user, so the identity is accurate by
+ * construction without peer credentials.
+ */
+export function hostUser(): string {
+  try {
+    return os.userInfo().username;
+    // eslint-disable-next-line no-catch-all/no-catch-all -- os.userInfo throws on exotic hosts; a fallback actor id beats no audit event
+  } catch {
+    return process.env.USER || 'unknown';
+  }
+}
+
+export function containerOrigin(sessionId: string, messagingGroupId: string | null): AuditOrigin {
+  const origin: AuditOrigin = { transport: 'container', session_id: sessionId };
+  if (messagingGroupId) {
+    origin.messaging_group_id = messagingGroupId;
+    const channel = getMessagingGroup(messagingGroupId)?.channel_type;
+    if (channel) origin.channel = channel;
+  }
+  return origin;
+}

.claude/skills/add-audit/add/src/cli/dispatch.audit.test.ts

@@ -0,0 +1,297 @@
+/**
+ * Audit middleware behavior of the exported dispatch — what gets recorded,
+ * for whom, and how gated chains correlate. Drives the real wrapped dispatch
+ * (real registry, real guard); audit is force-enabled and the store's append
+ * is captured. DB reads and approval delivery are mocked.
+ */
+import os from 'os';
+import { beforeEach, describe, expect, it, vi } from 'vitest';
+
+import type { PendingApproval } from '../types.js';
+
+const appended = vi.hoisted(() => ({ lines: [] as string[] }));
+const pendingRows = vi.hoisted(() => ({ rows: [] as unknown[] }));
+
+vi.mock('../audit/config.js', () => ({
+  AUDIT_ENABLED: true,
+  AUDIT_RETENTION_DAYS: 90,
+}));
+
+// Neutralize the adapter's module-scope boot (writability assert, prune,
+// maintenance timer) — the middleware is the unit under test here.
+vi.mock('../audit/init.js', () => ({
+  initAuditLog: vi.fn(),
+  maintainAudit: vi.fn(),
+}));
+
+vi.mock('../audit/store.js', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('../audit/store.js')>();
+  return {
+    ...actual,
+    appendAuditLine: (line: string) => {
+      appended.lines.push(line);
+    },
+  };
+});
+
+vi.mock('../log.js', () => ({
+  log: { debug: vi.fn(), info: vi.fn(), warn: vi.fn(), error: vi.fn(), fatal: vi.fn() },
+}));
+
+const mockGetContainerConfig = vi.fn();
+vi.mock('../db/container-configs.js', () => ({
+  getContainerConfig: (...args: unknown[]) => mockGetContainerConfig(...args),
+}));
+
+vi.mock('../db/agent-groups.js', () => ({
+  getAgentGroup: vi.fn(() => ({ id: 'g1', name: 'Group One' })),
+}));
+
+const mockGetPendingApproval = vi.fn();
+vi.mock('../db/sessions.js', () => ({
+  getSession: vi.fn(() => ({ id: 's1', agent_group_id: 'g1', messaging_group_id: 'mg1' })),
+  getPendingApproval: (...args: unknown[]) => mockGetPendingApproval(...args),
+  getPendingApprovalsByAction: () => pendingRows.rows,
+}));
+
+vi.mock('../db/messaging-groups.js', () => ({
+  getMessagingGroup: vi.fn(() => ({ channel_type: 'slack' })),
+}));
+
+const mockGetResource = vi.fn();
+vi.mock('./crud.js', () => ({
+  getResource: (...args: unknown[]) => mockGetResource(...args),
+}));
+
+vi.mock('../modules/approvals/index.js', () => ({
+  registerApprovalHandler: vi.fn(),
+  requestApproval: vi.fn(async () => undefined),
+}));
+
+import { register } from './registry.js';
+
+register({
+  name: 'groups-test',
+  description: 'echo command on the groups resource',
+  action: 'groups.test',
+  resource: 'groups',
+  access: 'open',
+  parseArgs: (raw) => raw,
+  handler: async (args) => ({ echo: args }),
+});
+
+register({
+  name: 'groups-get',
+  description: 'echo command for dash-joined id resolution',
+  action: 'groups.get',
+  resource: 'groups',
+  access: 'open',
+  parseArgs: (raw) => raw,
+  handler: async (args) => ({ echo: args }),
+});
+
+register({
+  name: 'wirings-list',
+  description: 'not on the group-scope allowlist',
+  action: 'wirings.list',
+  resource: 'wirings',
+  access: 'open',
+  parseArgs: (raw) => raw,
+  handler: async () => [],
+});
+
+register({
+  name: 'groups-fail',
+  description: 'handler that throws',
+  action: 'groups.fail',
+  resource: 'groups',
+  access: 'open',
+  parseArgs: (raw) => raw,
+  handler: async () => {
+    throw new Error('boom');
+  },
+});
+
+register({
+  name: 'groups-gated',
+  description: 'approval-gated command',
+  action: 'groups.gated',
+  resource: 'groups',
+  access: 'approval',
+  parseArgs: (raw) => raw,
+  handler: async () => 'ran',
+});
+
+import { dispatch } from './dispatch.js';
+import type { CallerContext } from './frame.js';
+
+const AGENT_CTX: CallerContext = { caller: 'agent', sessionId: 's1', agentGroupId: 'g1', messagingGroupId: 'mg1' };
+
+function grantRow(frameId: string, command: string): PendingApproval {
+  return {
+    approval_id: 'appr-123-abc',
+    session_id: 's1',
+    request_id: 'appr-123-abc',
+    action: 'cli_command',
+    payload: JSON.stringify({ frame: { id: frameId, command, args: {} }, callerContext: AGENT_CTX }),
+    created_at: new Date().toISOString(),
+    agent_group_id: 'g1',
+    channel_type: null,
+    platform_id: null,
+    platform_message_id: null,
+    expires_at: null,
+    status: 'pending',
+    title: 'CLI: groups-gated',
+    options_json: '[]',
+    approver_user_id: null,
+  };
+}
+
+function events(): Array<Record<string, any>> {
+  return appended.lines.map((l) => JSON.parse(l));
+}
+
+beforeEach(() => {
+  vi.clearAllMocks();
+  appended.lines.length = 0;
+  pendingRows.rows = [];
+  mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+  mockGetResource.mockImplementation((plural: string) => (plural === 'groups' ? { scopeField: 'id' } : undefined));
+});
+
+describe('withAudit(dispatch)', () => {
+  it('records a success event for a host caller with socket origin and host actor', async () => {
+    const resp = await dispatch({ id: '1', command: 'groups-test', args: { foo: 'bar' } }, { caller: 'host' });
+
+    expect(resp.ok).toBe(true);
+    const [event] = events();
+    expect(event).toMatchObject({
+      schema_version: 1,
+      actor: { type: 'human', id: `host:${os.userInfo().username}`, email: null },
+      origin: { transport: 'socket' },
+      action: 'groups.test',
+      outcome: 'success',
+      correlation_id: null,
+      details: { foo: 'bar' },
+    });
+  });
+
+  it('records effective args after group auto-fill, with container origin and channel', async () => {
+    await dispatch({ id: '1', command: 'groups-test', args: {} }, AGENT_CTX);
+
+    const [event] = events();
+    expect(event.actor).toMatchObject({ type: 'agent', id: 'g1' });
+    expect(event.origin).toEqual({
+      transport: 'container',
+      session_id: 's1',
+      messaging_group_id: 'mg1',
+      channel: 'slack',
+    });
+    expect(event.details).toMatchObject({ id: 'g1', agent_group_id: 'g1', group: 'g1' });
+    expect(event.resources).toContainEqual({ type: 'agent_group', id: 'g1' });
+  });
+
+  it('records a denied event for a scope denial, naming the attempted resource type', async () => {
+    const resp = await dispatch({ id: '1', command: 'wirings-list', args: {} }, AGENT_CTX);
+
+    expect(resp.ok).toBe(false);
+    const [event] = events();
+    expect(event).toMatchObject({
+      action: 'wirings.list',
+      outcome: 'denied',
+      resources: [{ type: 'wirings' }],
+      details: { error: 'forbidden' },
+    });
+    expect(event.details.reason).toContain('scoped');
+  });
+
+  it('records a failure event when the handler throws', async () => {
+    await dispatch({ id: '1', command: 'groups-fail', args: {} }, { caller: 'host' });
+
+    const [event] = events();
+    expect(event).toMatchObject({ action: 'groups.fail', outcome: 'failure', details: { error: 'handler-error' } });
+    expect(event.details.reason).toContain('boom');
+  });
+
+  it('records a hold as a pending event correlated to the approval row it created', async () => {
+    pendingRows.rows = [grantRow('1', 'groups-gated')];
+
+    const resp = await dispatch({ id: '1', command: 'groups-gated', args: {} }, AGENT_CTX);
+
+    expect(resp.ok).toBe(false);
+    if (!resp.ok) expect(resp.error.code).toBe('approval-pending');
+    const [event] = events();
+    expect(event).toMatchObject({
+      action: 'groups.gated',
+      outcome: 'pending',
+      correlation_id: 'appr-123-abc',
+    });
+    expect(event.resources).toContainEqual({ type: 'approval', id: 'appr-123-abc' });
+    expect(event.details.error).toBeUndefined();
+  });
+
+  it('records an uncorrelated pending event when no approval row was created (no approver)', async () => {
+    pendingRows.rows = [];
+
+    await dispatch({ id: '1', command: 'groups-gated', args: {} }, AGENT_CTX);
+
+    const [event] = events();
+    expect(event).toMatchObject({ outcome: 'pending', correlation_id: null });
+  });
+
+  it('records an approved replay as success with the grant approval id as correlation_id', async () => {
+    const grant = grantRow('9', 'groups-gated');
+    mockGetPendingApproval.mockReturnValue(grant);
+
+    const resp = await dispatch({ id: '9', command: 'groups-gated', args: {} }, AGENT_CTX, { grant });
+
+    expect(resp.ok).toBe(true);
+    const [event] = events();
+    expect(event).toMatchObject({
+      action: 'groups.gated',
+      outcome: 'success',
+      correlation_id: 'appr-123-abc',
+    });
+    expect(event.resources).toContainEqual({ type: 'approval', id: 'appr-123-abc' });
+  });
+
+  it('records unknown commands as cli.unknown-command with the raw name in details', async () => {
+    await dispatch({ id: '1', command: 'nope-nothing', args: {} }, { caller: 'host' });
+
+    const [event] = events();
+    expect(event).toMatchObject({
+      action: 'cli.unknown-command',
+      outcome: 'failure',
+      resources: [],
+      details: { command: 'nope-nothing', error: 'unknown-command' },
+    });
+  });
+
+  it('records the resolved command and id for dash-joined positional ids', async () => {
+    const uuid = '550e8400-e29b-41d4-a716-446655440000';
+    await dispatch({ id: '1', command: `groups-get-${uuid}`, args: {} }, { caller: 'host' });
+
+    const [event] = events();
+    expect(event).toMatchObject({ action: 'groups.get', outcome: 'success' });
+    expect(event.resources).toContainEqual({ type: 'agent_group', id: uuid });
+    expect(event.details.id).toBe(uuid);
+  });
+
+  it('normalizes hyphenated arg keys in details', async () => {
+    await dispatch({ id: '1', command: 'groups-test', args: { 'dry-run': 'true' } }, { caller: 'host' });
+
+    const [event] = events();
+    expect(event.details).toMatchObject({ dry_run: 'true' });
+  });
+
+  it('parses JSON string args so the redactor reaches their inner keys', async () => {
+    await dispatch(
+      { id: '1', command: 'groups-test', args: { env: '{"NOTION_TOKEN":"tok-123","SAFE":"ok"}', note: '{not json' } },
+      { caller: 'host' },
+    );
+
+    const [event] = events();
+    expect(event.details.env).toEqual({ NOTION_TOKEN: '[REDACTED]', SAFE: 'ok' });
+    expect(event.details.note).toBe('{not json');
+  });
+});

.claude/skills/add-audit/add/src/cli/dispatch.audit.ts

@@ -0,0 +1,238 @@
+/**
+ * CLI audit adapter (installed by /add-audit) — owns how the dispatcher
+ * describes itself to the audit log: the dispatch middleware plus the
+ * CLI-specific actor/origin/resource mapping. Composed in dispatch.ts as
+ * `export const dispatch = withAudit(dispatchInner)`; business logic there
+ * contains zero audit calls.
+ *
+ * Loading this module also boots the audit log (writability assert, boot
+ * prune, hook lifecycle, maintenance timer): dispatch.ts is imported by both
+ * transports during the host's barrel phase, so initAuditLog() runs before
+ * any command is accepted — and an enabled box with an unwritable
+ * data/audit/ refuses to start.
+ */
+import { emitAuditEvent } from '../audit/emit.js';
+import { initAuditLog } from '../audit/init.js';
+import { type AuditActor, type AuditOrigin, type AuditOutcome, type AuditResource } from '../audit/types.js';
+import { containerOrigin, hostUser } from '../audit/vocab.js';
+import { getContainerConfig } from '../db/container-configs.js';
+import { getPendingApprovalsByAction } from '../db/sessions.js';
+import type { PendingApproval } from '../types.js';
+import { getResource } from './crud.js';
+import type { CallerContext, RequestFrame, ResponseFrame } from './frame.js';
+import { type CommandDef, lookup } from './registry.js';
+
+initAuditLog();
+
+// ── CLI mapping ──
+
+/**
+ * Host callers stamp `host:<install user>` daemon-side (the ncl socket is
+ * 0600 and owned by the install user); container callers are their agent group.
+ */
+export function actorForCaller(ctx: CallerContext): AuditActor {
+  return ctx.caller === 'host' ? { type: 'human', id: `host:${hostUser()}` } : { type: 'agent', id: ctx.agentGroupId };
+}
+
+export function originForCaller(ctx: CallerContext): AuditOrigin {
+  if (ctx.caller === 'host') return { transport: 'socket' };
+  return containerOrigin(ctx.sessionId, ctx.messagingGroupId || null);
+}
+
+/**
+ * Frame-level args use `--hyphen-keys`; recorded details use the same
+ * underscore form the parsed handlers see. Mirrors crud's normalizeArgs
+ * (kept local so audit doesn't depend on a module tests commonly mock).
+ */
+export function normalizeArgKeys(raw: Record<string, unknown>): Record<string, unknown> {
+  const out: Record<string, unknown> = {};
+  for (const [k, v] of Object.entries(raw)) {
+    out[k.replace(/-/g, '_')] = v;
+  }
+  return out;
+}
+
+/** CLI resource plural → audit resource type, where the singular isn't it. */
+const RESOURCE_TYPE_OVERRIDES: Record<string, string> = {
+  groups: 'agent_group',
+  'messaging-groups': 'messaging_group',
+  'dropped-messages': 'dropped_message',
+  'user-dms': 'user_dm',
+};
+
+/**
+ * Derive touched/attempted resources from a command's effective args. Generic
+ * by design: `id` → the command's own resource, group/user args → their
+ * types, and a bare `{type}` entry when nothing else is known (a denied
+ * `users list` still names what was attempted).
+ */
+export function resourcesForCli(cmd: CommandDef, args: Record<string, unknown>): AuditResource[] {
+  if (!cmd.resource) return [];
+  const type = RESOURCE_TYPE_OVERRIDES[cmd.resource] ?? getResource(cmd.resource)?.name ?? cmd.resource;
+
+  const out: AuditResource[] = [];
+  const push = (t: string, id: unknown): void => {
+    if (typeof id !== 'string' || !id) return;
+    if (!out.some((r) => r.type === t && r.id === id)) out.push({ type: t, id });
+  };
+  push(type, args.id);
+  push('agent_group', args.agent_group_id ?? args.group);
+  push('user', args.user);
+  if (out.length === 0) out.push({ type });
+  return out;
+}
+
+// ── Dispatch mechanics, mirrored for the record ──
+// Dispatch resolves the command and auto-fills args on a local copy of the
+// frame that never leaves it, so the middleware mirrors the two documented
+// mechanics below. The copies are mechanical, and drift only ever degrades a
+// record's detail (a fallback action name, a missing auto-filled arg) — never
+// dispatch behavior, and never an outcome.
+
+/**
+ * Mirror of dispatch's command resolution: exact lookup, then the longest
+ * registered dash-prefix with the remainder recorded as --id.
+ */
+function resolveForRecord(req: RequestFrame): { cmd?: CommandDef; args: Record<string, unknown> } {
+  const direct = lookup(req.command);
+  if (direct) return { cmd: direct, args: req.args };
+  let shortened = req.command;
+  let idx: number;
+  while ((idx = shortened.lastIndexOf('-')) > 0) {
+    shortened = shortened.slice(0, idx);
+    const fallback = lookup(shortened);
+    if (fallback) {
+      const tail = req.command.slice(shortened.length + 1);
+      return { cmd: fallback, args: { ...req.args, id: req.args.id ?? tail } };
+    }
+  }
+  return { args: req.args };
+}
+
+/**
+ * Mirror of dispatch's group-scope arg auto-fill, so the record shows the
+ * effective args the handler saw (e.g. which agent group a bare
+ * `sessions list` actually listed).
+ */
+function effectiveArgs(
+  cmd: CommandDef | undefined,
+  args: Record<string, unknown>,
+  ctx: CallerContext,
+): Record<string, unknown> {
+  if (ctx.caller !== 'agent') return args;
+  if ((getContainerConfig(ctx.agentGroupId)?.cli_scope ?? 'group') !== 'group') return args;
+  const fill: Record<string, unknown> = {
+    agent_group_id: args.agent_group_id ?? ctx.agentGroupId,
+    group: args.group ?? ctx.agentGroupId,
+  };
+  if (cmd?.resource === 'groups' || cmd?.resource === 'destinations') {
+    fill.id = args.id ?? ctx.agentGroupId;
+  }
+  return { ...args, ...fill };
+}
+
+/**
+ * CLI args arrive as strings, so a JSON-object/array value (e.g.
+ * `--env '{"NOTION_TOKEN":"…"}'`) would reach the redactor as one opaque
+ * string under an innocent key and its inner secret keys would survive.
+ * Recording the parsed form lets the redactor walk inside — the audit log's
+ * "secrets never land" property depends on it for exactly this flow.
+ */
+function parseJsonishValues(args: Record<string, unknown>): Record<string, unknown> {
+  const out: Record<string, unknown> = {};
+  for (const [k, v] of Object.entries(args)) {
+    if (typeof v === 'string' && (v.startsWith('{') || v.startsWith('['))) {
+      try {
+        out[k] = JSON.parse(v);
+        continue;
+        // eslint-disable-next-line no-catch-all/no-catch-all -- not JSON after all: record the string as-is
+      } catch {
+        /* fall through */
+      }
+    }
+    out[k] = v;
+  }
+  return out;
+}
+
+/**
+ * The approval row a hold just created for this frame — it gives the pending
+ * event the same correlation_id the approved replay will carry as its guard
+ * grant. requestApproval keeps the minted id internal, so the row is
+ * recovered by the frame id it stored in its payload; no row (e.g. no
+ * configured approver) → the hold is still recorded, uncorrelated.
+ */
+function holdApprovalIdFor(frameId: string): string | null {
+  const rows = getPendingApprovalsByAction('cli_command');
+  for (let i = rows.length - 1; i >= 0; i--) {
+    try {
+      const payload = JSON.parse(rows[i].payload) as { frame?: { id?: string } };
+      if (payload.frame?.id === frameId) return rows[i].approval_id;
+      // eslint-disable-next-line no-catch-all/no-catch-all -- a row with an unparseable payload is simply not this frame's hold
+    } catch {
+      continue;
+    }
+  }
+  return null;
+}
+
+// ── The dispatch middleware ──
+
+type DispatchInner = (
+  req: RequestFrame,
+  ctx: CallerContext,
+  opts?: { grant?: PendingApproval },
+) => Promise<ResponseFrame>;
+
+/**
+ * Dispatch middleware — the exported `dispatch` is the wrapped function, so
+ * the socket server, the container delivery-action, and the in-module
+ * approved replay are all covered by the one composition.
+ *
+ * Outcome derives from the response frame: ok → success, forbidden → denied
+ * (captures pre-handler scope denials), approval-pending → pending (the
+ * record of a hold), anything else → failure. Correlation is the approval id:
+ * an approved replay carries the approval row as its guard grant
+ * (opts.grant), and a fresh hold is correlated by recovering the row it just
+ * created — so `--correlation <approval-id>` returns the whole gated chain.
+ */
+export function withAudit(inner: DispatchInner): DispatchInner {
+  return async (req, ctx, opts = {}) => {
+    const res = await inner(req, ctx, opts);
+    emitAuditEvent(() => {
+      const resolved = resolveForRecord(req);
+      const cmd = resolved.cmd;
+      const pending = !res.ok && res.error.code === 'approval-pending';
+      const outcome: AuditOutcome = res.ok
+        ? 'success'
+        : res.error.code === 'forbidden'
+          ? 'denied'
+          : pending
+            ? 'pending'
+            : 'failure';
+      // A denial records the attempt as asked (raw args): nothing ran, so the
+      // auto-fill never conceptually happened — and a cross-group attempt
+      // must show the foreign id the caller passed, not a filled-in own-group.
+      const args = normalizeArgKeys(outcome === 'denied' ? resolved.args : effectiveArgs(cmd, resolved.args, ctx));
+      const correlationId = opts.grant?.approval_id ?? (pending ? holdApprovalIdFor(req.id) : null);
+      const details: Record<string, unknown> = parseJsonishValues(args);
+      if (!res.ok && !pending) {
+        details.error = res.error.code;
+        details.reason = res.error.message;
+      }
+      if (!cmd) details.command = req.command;
+      const resources = cmd ? resourcesForCli(cmd, args) : [];
+      if (correlationId) resources.push({ type: 'approval', id: correlationId });
+      return {
+        actor: actorForCaller(ctx),
+        origin: originForCaller(ctx),
+        action: cmd ? (cmd.action ?? `cli.${cmd.name}`) : 'cli.unknown-command',
+        resources,
+        outcome,
+        correlationId,
+        details,
+      };
+    });
+    return res;
+  };
+}

.claude/skills/add-audit/add/src/cli/resources/audit.ts

@@ -0,0 +1,56 @@
+import { listAuditEvents } from '../../audit/reader.js';
+import { registerResource } from '../crud.js';
+
+/**
+ * Read-only audit resource (installed by /add-audit). The "table" is the
+ * NDJSON day-file store — no generic CRUD verbs are declared, so it is never
+ * queried as SQL; `list` is a custom operation backed by the audit reader.
+ *
+ * Deliberately NOT on the group-scope allowlist (GROUP_SCOPE_RESOURCES):
+ * audit spans agent groups, so group-scoped agents are refused before the
+ * handler — the resource is host + `cli_scope: global` only, by omission
+ * (fails closed).
+ */
+registerResource({
+  name: 'audit_event',
+  plural: 'audit',
+  table: '(data/audit/*.ndjson)',
+  description: 'Local audit log — one event per ncl command. Newest first. Requires AUDIT_ENABLED=true.',
+  idColumn: 'event_id',
+  columns: [
+    { name: 'actor', type: 'string', description: 'Filter: exact actor id (host:<user>, <channel>:<handle>, agent group id)' },
+    { name: 'action', type: 'string', description: 'Filter: exact action or dotted prefix (e.g. groups.config)' },
+    { name: 'resource', type: 'string', description: 'Filter: matches any event resource by id or type' },
+    {
+      name: 'outcome',
+      type: 'string',
+      description: 'Filter: event outcome',
+      enum: ['success', 'failure', 'denied', 'pending', 'approved', 'rejected'],
+    },
+    { name: 'since', type: 'string', description: 'Window start: 7d / 24h / 30m relative, or ISO date' },
+    { name: 'until', type: 'string', description: 'Window end: same formats as --since' },
+    { name: 'correlation', type: 'string', description: 'Filter: approval id tying a gated chain together' },
+    { name: 'limit', type: 'number', description: 'Max events (default 100, newest first)' },
+    { name: 'format', type: 'string', description: 'Output format', enum: ['ndjson'] },
+  ],
+  operations: {},
+  customOperations: {
+    list: {
+      access: 'open',
+      description: 'Query audit events, newest first. --format ndjson streams the stored lines for SIEM export.',
+      examples: [
+        'ncl audit list --outcome denied --since 7d',
+        'ncl audit list --correlation appr-1751970000000-x1y2z3',
+        'ncl audit list --action groups.config --format ndjson',
+      ],
+      handler: async (args) => listAuditEvents(args),
+      formatHuman: (data) => {
+        // NDJSON export prints the stored lines verbatim; for the table view
+        // this throws so dispatch falls back to the row objects, which every
+        // client already renders.
+        if (typeof data === 'string') return data;
+        throw new Error('render rows client-side');
+      },
+    },
+  },
+});

CHANGELOG.md

@@ -4,6 +4,9 @@ All notable changes to NanoClaw will be documented in this file.
 
 ## [Unreleased]
 
+- **New skill: `/add-audit` — opt-in local audit log for the `ncl` surface.** Installs an append-only, SIEM-shaped audit log: every command through the dispatcher (both transports — host socket and container — including scope denials, approval holds, and approved replays) becomes one canonical NDJSON event under `data/audit/<UTC-day>.ndjson`. Gated chains share a `correlation_id` (the approval id: the hold's `pending` event and the replay's terminal event carry the same one); `details` pass a recursive secret-key redactor (~2 KB value cap). `AUDIT_RETENTION_DAYS` (default 90; `0` = keep forever) hard-deletes day-files past the horizon at boot and once per UTC day. Read back with `ncl audit list [--actor --action --resource --outcome --since --until --correlation --limit] [--format ndjson]` — host + global-scope callers only (the resource stays off the group-scope allowlist, so group-scoped agents fail closed). In-process exporters plug in via `registerAuditHook` — post-write hooks that fire only after the local append succeeds. Off by default: nothing is persisted (and `data/audit/` is never created) until `AUDIT_ENABLED=true`, an enabled box refuses to boot if the directory isn't writable, and a disabled box answers `ncl audit list` with a clear error instead of an empty list. The skill's whole core footprint is the dispatch composition (`export const dispatch = withAudit(dispatchInner)`) plus the resource-barrel import, both guarded by a shipped wiring test. Approval-lifecycle events (request/decision as their own events), OneCLI credential holds, and channel/sender events are later increments on the same schema.
+- **The guard seam (guarded-actions phase 2).** Every privileged action crossing the container or channel boundary now passes ONE decision function — `guard()` in the new `src/guard/` leaf — before it executes: `allow`, `hold` (through the flow's existing approval mechanics), or `deny`. The four handler registries (ncl commands, delivery actions, response handlers, message interceptors) wrap their handlers at registration, so the guarded path is the only path by construction, and a registry-walking conformance test fails CI on any unmapped mutating entry. The structural baseline is today's checks verbatim (cli_scope enforcement, create_agent's scope branch, a2a destination ACL + `agent_message_policies` hold, self-mod unconditional hold, unknown_sender_policy, channel-registration click auth, host-as-trusted-caller in code). Policy-as-data (tighten-only rule sources composing with the baseline) is deferred to phase 3, where the generalized rules table arrives with its first operator-visible consumer. Approved replays now carry the verified approval row as a **grant** — the `approved: true` boolean is deleted — and the guard re-runs the structural baseline on every replay. Three outcome changes are inherent to that replay semantics and ship deliberately: **(a)** approving a held a2a message after its destination was revoked no longer delivers it (the live baseline re-check); **(b)** a forged, already-consumed, or wrong-action/wrong-payload grant refuses the replay instead of executing (previously any in-process caller passing `approved: true` executed); **(c)** the channel-registration free-text name reply re-checks approver eligibility at reply time — a privilege revoked between the click and the reply, or a registration that vanished meanwhile, now refuses instead of acting. Zero rules ship by default — the seeded posture is today's behavior. (The D1/D2/D4 click-auth fixes from the guarded-actions defect inventory are NOT in this change — they belong to the approval-contract PR.)
+- **Guard conformance is now a boot invariant, not just a CI test.** The host refuses to start (upgrade-tripwire posture: clear banner + exit 1) if any privileged command or delivery action is registered without a guard-catalog mapping. CI can't see skill-installed code — `/add-*` skills register handlers on machines where the test suite never runs — so the same registry walk the conformance test uses (`src/guard-conformance.ts`, one shared exemption list) now runs in `main()` before the host accepts a message, and a bad registration crashes at skill-install time instead of running unguarded. Companion hardening: re-registering a guard-wrapped delivery action *without* a guard spec now throws at registration — previously that silently replaced the guarded handler while the catalog still reported the action as guarded.
 - **Optional per-container resource caps.** `CONTAINER_CPU_LIMIT` and `CONTAINER_MEMORY_LIMIT` pass through to `docker run` as `--cpus` / `--memory` (`container-runner.ts`). Both empty by default — no flag added, spawn args byte-identical to today — so existing installs are unaffected. Set them to cap an agent container's CPU/memory so one agent can't monopolize the host (e.g. `CONTAINER_CPU_LIMIT=2`, `CONTAINER_MEMORY_LIMIT=8g`). Swap is intentionally not managed here: `--memory` is a hard cap on a swapless host.
 - [BREAKING] **Chat SDK pinned to `4.29.0` (was `4.26.0` via `^4.24.0`).** `chat` and the `@chat-adapter/*` channel adapters are version-locked — the adapter's `ChatInstance` must match the bridge's, so a mismatched pair fails to typecheck at `createChatSdkBridge(...)`. `chat` is therefore pinned exactly, and the channel-adapter install pins move with it — the `/add-<channel>` SKILL.md steps and `setup/*.sh` scripts on `main`, plus the adapter code on the `channels` branch. Core installs with no channel (only `cli`) are unaffected. **Migration:** if any channel is installed (Slack, Discord, Telegram, Teams, …), re-run its `/add-<channel>` skill to pull the matching `4.29.0` adapter.
 - **Budget/billing-exhausted LLM turns now reach the user instead of being silently dropped.** When a turn ends in a non-retryable provider error (e.g. an Anthropic `403 billing_error`) with no `<message>` wrapping, the agent-runner delivers the provider's notice to the originating channel and stops re-nudging the failing gateway. `providers/claude.ts` now surfaces the SDK's `is_error` flag (and the error subtype's `errors[]` text); `poll-loop.ts` delivers that text and skips the re-wrap retry. Fixes the case where a spend-limit notice produced silence plus a turn-after-turn retry loop.

CLAUDE.md

@@ -66,6 +66,7 @@ For ad-hoc queries from skills or scripts, use the in-tree wrapper rather than t
 | `src/session-manager.ts` | Resolves sessions; opens `inbound.db` / `outbound.db`; manages heartbeat path |
 | `src/container-runner.ts` | Spawns per-agent-group Docker containers with session DB + outbox mounts, OneCLI `ensureAgent` |
 | `src/container-runtime.ts` | Docker CLI wrapper (runtime binary, host-gateway args, mount args), orphan cleanup |
+| `src/guard/` | Privileged-action decision seam: `guard(action, input)` → allow \| hold \| deny. Domain-free leaf; domain baselines are `defineGuardedAction` values exported from module-edge `guard.ts` adapters (cli, agent-to-agent, self-mod, permissions) — consults hold the value, so a missing/typo'd wiring is a compile error, never a fail-open. All four handler registries require a guard spec or an explicit `unguarded(<reason>)` declaration at the registration site. Approved replays carry the approval row as a grant and re-check the structural baseline. Boot check = grant continuations only (`src/guard-conformance.ts`); Policy-as-data is deferred to guarded-actions phase 3. Conformance test: `src/guard/conformance.test.ts` |
 | `src/modules/permissions/access.ts` | `canAccessAgentGroup` — owner / global admin / scoped admin / member resolution against `user_roles` + `agent_group_members` |
 | `src/modules/approvals/primitive.ts` | `pickApprover`, `pickApprovalDelivery`, `requestApproval`, approval-handler registry |
 | `src/command-gate.ts` | Router-side admin command gate — queries `user_roles` directly (no env var, no container-side check) |

src/cli/crud.ts

@@ -327,6 +327,7 @@ export function registerResource(def: ResourceDef): void {
   if (def.operations.list) {
     register({
       name: `${def.plural}-list`,
+      action: `${def.plural}.list`,
       description: `List all ${def.plural}.`,
       access: def.operations.list,
       resource: def.plural,
@@ -339,6 +340,7 @@ export function registerResource(def: ResourceDef): void {
   if (def.operations.get) {
     register({
       name: `${def.plural}-get`,
+      action: `${def.plural}.get`,
       description: `Get a ${def.name} by ID.`,
       access: def.operations.get,
       resource: def.plural,
@@ -351,6 +353,7 @@ export function registerResource(def: ResourceDef): void {
   if (def.operations.create) {
     register({
       name: `${def.plural}-create`,
+      action: `${def.plural}.create`,
       description: `Create a new ${def.name}.`,
       access: def.operations.create,
       resource: def.plural,
@@ -362,6 +365,7 @@ export function registerResource(def: ResourceDef): void {
   if (def.operations.update) {
     register({
       name: `${def.plural}-update`,
+      action: `${def.plural}.update`,
       description: `Update a ${def.name}.`,
       access: def.operations.update,
       resource: def.plural,
@@ -373,6 +377,7 @@ export function registerResource(def: ResourceDef): void {
   if (def.operations.delete) {
     register({
       name: `${def.plural}-delete`,
+      action: `${def.plural}.delete`,
       description: `Delete a ${def.name}.`,
       access: def.operations.delete,
       resource: def.plural,
@@ -389,6 +394,7 @@ export function registerResource(def: ResourceDef): void {
       const declared = op.args;
       register({
         name: `${def.plural}-${verb.replace(/ /g, '-')}`,
+        action: `${def.plural}.${verb.replace(/ /g, '.')}`,
         description: op.description,
         access: op.access,
         resource: def.plural,

src/cli/delivery-action.ts

@@ -8,52 +8,57 @@
 import type Database from 'better-sqlite3';
 
 import { registerDeliveryAction } from '../delivery.js';
+import { unguarded } from '../guard/index.js';
 import { insertMessage } from '../db/session-db.js';
 import { log } from '../log.js';
 import { dispatch } from './dispatch.js';
 import type { RequestFrame } from './frame.js';
 import type { Session } from '../types.js';
 
-registerDeliveryAction('cli_request', async (content, session, inDb) => {
-  const requestId = content.requestId as string;
-  const command = content.command as string;
-  const args = (content.args as Record<string, unknown>) ?? {};
+registerDeliveryAction(
+  'cli_request',
+  async (content, session, inDb) => {
+    const requestId = content.requestId as string;
+    const command = content.command as string;
+    const args = (content.args as Record<string, unknown>) ?? {};
 
-  if (!requestId || !command) {
-    log.warn('cli_request missing requestId or command', { sessionId: session.id });
-    return;
-  }
+    if (!requestId || !command) {
+      log.warn('cli_request missing requestId or command', { sessionId: session.id });
+      return;
+    }
 
-  const req: RequestFrame = { id: requestId, command, args };
-  const ctx = {
-    caller: 'agent' as const,
-    sessionId: session.id,
-    agentGroupId: session.agent_group_id,
-    messagingGroupId: session.messaging_group_id ?? '',
-  };
+    const req: RequestFrame = { id: requestId, command, args };
+    const ctx = {
+      caller: 'agent' as const,
+      sessionId: session.id,
+      agentGroupId: session.agent_group_id,
+      messagingGroupId: session.messaging_group_id ?? '',
+    };
 
-  log.info('CLI request from agent', { requestId, command, sessionId: session.id });
+    log.info('CLI request from agent', { requestId, command, sessionId: session.id });
 
-  const response = await dispatch(req, ctx);
+    const response = await dispatch(req, ctx);
 
-  // Write response to inbound.db so the container can read it.
-  // trigger=0: don't wake the agent — this is an inline response to a tool call.
-  insertMessage(inDb, {
-    id: `cli-resp-${requestId}`,
-    kind: 'system',
-    timestamp: new Date().toISOString(),
-    platformId: null,
-    channelType: null,
-    threadId: null,
-    content: JSON.stringify({
-      type: 'cli_response',
-      requestId,
-      frame: response,
-    }),
-    processAfter: null,
-    recurrence: null,
-    trigger: 0,
-  });
+    // Write response to inbound.db so the container can read it.
+    // trigger=0: don't wake the agent — this is an inline response to a tool call.
+    insertMessage(inDb, {
+      id: `cli-resp-${requestId}`,
+      kind: 'system',
+      timestamp: new Date().toISOString(),
+      platformId: null,
+      channelType: null,
+      threadId: null,
+      content: JSON.stringify({
+        type: 'cli_response',
+        requestId,
+        frame: response,
+      }),
+      processAfter: null,
+      recurrence: null,
+      trigger: 0,
+    });
 
-  log.info('CLI response written', { requestId, ok: response.ok, sessionId: session.id });
-});
+    log.info('CLI response written', { requestId, ok: response.ok, sessionId: session.id });
+  },
+  unguarded('transport envelope — every inner command is guarded at dispatch'),
+);

src/cli/dispatch.test.ts

@@ -9,6 +9,7 @@ const approvalState = vi.hoisted(() => ({
     | ((args: {
         session: unknown;
         payload: Record<string, unknown>;
+        approval: Record<string, unknown>;
         userId: string;
         notify: (text: string) => void;
       }) => Promise<void>),
@@ -18,6 +19,7 @@ const approvalState = vi.hoisted(() => ({
       handler: (args: {
         session: unknown;
         payload: Record<string, unknown>;
+        approval: Record<string, unknown>;
         userId: string;
         notify: (text: string) => void;
       }) => Promise<void>,
@@ -43,8 +45,11 @@ vi.mock('../db/agent-groups.js', () => ({
 }));
 
 const mockGetSession = vi.fn();
+// The guard's grant check re-fetches the approval row to prove it's live.
+const mockGetPendingApproval = vi.fn();
 vi.mock('../db/sessions.js', () => ({
   getSession: (...args: unknown[]) => mockGetSession(...args),
+  getPendingApproval: (...args: unknown[]) => mockGetPendingApproval(...args),
 }));
 
 // dispatch's post-handler looks up the resource's `scopeField` via getResource.
@@ -472,10 +477,20 @@ describe('CLI scope enforcement', () => {
       callerContext: ctx,
     });
 
+    // The approve path hands the handler the live approval row — the grant
+    // the replay carries back into dispatch.
+    const grantRow = {
+      approval_id: 'appr-t1',
+      action: 'cli_command',
+      payload: JSON.stringify(approval.payload),
+    };
+    mockGetPendingApproval.mockReturnValue(grantRow);
+
     expect(approvalState.approvalHandler).toBeTypeOf('function');
     await approvalState.approvalHandler!({
       session: { id: 's1', agent_group_id: 'g1', messaging_group_id: 'mg1' },
       payload: approval.payload,
+      approval: grantRow,
       userId: 'telegram:admin',
       notify: vi.fn(),
     });
@@ -484,6 +499,73 @@ describe('CLI scope enforcement', () => {
     expect(approvalState.requestApproval).toHaveBeenCalledTimes(1);
   });
 
+  // --- Grant-carrying replay (the `approved: true` boolean no longer exists) ---
+
+  it('replay with a dead grant (row deleted) refuses instead of re-holding', async () => {
+    mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+    mockGetSession.mockReturnValue({ id: 's1', agent_group_id: 'g1', messaging_group_id: 'mg1' });
+    mockGetAgentGroup.mockReturnValue({ id: 'g1', name: 'Group One' });
+
+    const ctx = agentCtx();
+    await dispatch({ id: '1', command: 'approval-context-command', args: {} }, ctx);
+    const approval = approvalState.requestApproval.mock.calls[0][0] as { payload: Record<string, unknown> };
+
+    mockGetPendingApproval.mockReturnValue(undefined); // resolution already deleted the row
+    const notify = vi.fn();
+    await approvalState.approvalHandler!({
+      session: { id: 's1', agent_group_id: 'g1', messaging_group_id: 'mg1' },
+      payload: approval.payload,
+      approval: { approval_id: 'appr-dead', action: 'cli_command', payload: JSON.stringify(approval.payload) },
+      userId: 'telegram:admin',
+      notify,
+    });
+
+    expect(approvalState.observedContexts).toHaveLength(0); // handler never ran
+    expect(approvalState.requestApproval).toHaveBeenCalledTimes(1); // no second card
+    expect(notify.mock.calls[0][0]).toContain('failed');
+  });
+
+  it("a grant approved for one command doesn't transfer to another", async () => {
+    mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+
+    // A live cli_command row, but held for a DIFFERENT command.
+    const grantRow = {
+      approval_id: 'appr-other',
+      action: 'cli_command',
+      payload: JSON.stringify({ frame: { command: 'members-add' } }),
+    };
+    mockGetPendingApproval.mockReturnValue(grantRow);
+
+    const resp = await dispatch({ id: '1', command: 'approval-context-command', args: {} }, agentCtx(), {
+      grant: grantRow as never,
+    });
+
+    expect(resp.ok).toBe(false);
+    if (!resp.ok) {
+      expect(resp.error.code).toBe('forbidden');
+      expect(resp.error.message).toContain('grant');
+    }
+    expect(approvalState.observedContexts).toHaveLength(0);
+  });
+
+  it('a fabricated grant object without a live row is refused', async () => {
+    mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+    mockGetPendingApproval.mockReturnValue(undefined);
+
+    const forged = {
+      approval_id: 'appr-forged',
+      action: 'cli_command',
+      payload: JSON.stringify({ frame: { command: 'approval-context-command' } }),
+    };
+    const resp = await dispatch({ id: '1', command: 'approval-context-command', args: {} }, agentCtx(), {
+      grant: forged as never,
+    });
+
+    expect(resp.ok).toBe(false);
+    if (!resp.ok) expect(resp.error.code).toBe('forbidden');
+    expect(approvalState.requestApproval).not.toHaveBeenCalled(); // refusal, not a fresh hold
+  });
+
   // --- Post-handler filtering ---
 
   it('group: groups list filters out other groups', async () => {

src/cli/dispatch.ts

@@ -3,23 +3,37 @@
  * the per-session DB poller (container caller) call dispatch() with the
  * same frame and a transport-supplied CallerContext.
  *
- * Approval gating for risky calls from the container is the only branch
- * that differs by caller. Host callers and `open` commands run inline.
+ * Every command passes the guard before its handler runs — the decision
+ * (allow / hold / deny) comes from the command's catalog entry, derived at
+ * registration (see cli/guard.ts). Dispatch keeps the mechanics: arg
+ * auto-fill, the sessions-get existence oracle, `--help` interception,
+ * parseArgs, and post-handler row filtering. An approved replay re-enters
+ * here carrying the verified approval row as its grant — the guard re-checks
+ * the structural baseline live, and the `approved: true` boolean no longer
+ * exists.
  */
 import { getContainerConfig } from '../db/container-configs.js';
 import { getAgentGroup } from '../db/agent-groups.js';
 import { getSession } from '../db/sessions.js';
+import { guard, type GuardActor } from '../guard/index.js';
 import { registerApprovalHandler, requestApproval } from '../modules/approvals/index.js';
+import type { PendingApproval } from '../types.js';
 import type { CallerContext, ErrorCode, RequestFrame, ResponseFrame } from './frame.js';
 import { getResource } from './crud.js';
 import { listVerbs, renderVerbHelp } from './help-render.js';
-import { GROUP_SCOPE_RESOURCES, listCommands, lookup } from './registry.js';
+import { commandGuard, listCommands, lookup } from './registry.js';
 
 type DispatchOptions = {
-  /** True when a command is being replayed after approval. */
-  approved?: boolean;
+  /** Verified approval row when a command is replayed after approval. */
+  grant?: PendingApproval;
 };
 
+function actorFor(ctx: CallerContext): GuardActor {
+  return ctx.caller === 'host'
+    ? { kind: 'host' }
+    : { kind: 'agent', agentGroupId: ctx.agentGroupId, sessionId: ctx.sessionId };
+}
+
 export async function dispatch(
   req: RequestFrame,
   ctx: CallerContext,
@@ -54,43 +68,13 @@ export async function dispatch(
     return err(req.id, 'unknown-command', unknownCommandMessage(req.command));
   }
 
-  // CLI scope enforcement for agent callers
+  // Group-scope mechanics for agent callers (visibility, not policy — the
+  // allow/hold/deny decisions live in the guard baseline, cli/guard.ts).
   if (ctx.caller === 'agent') {
     const configRow = getContainerConfig(ctx.agentGroupId);
     const cliScope = configRow?.cli_scope ?? 'group';
 
-    if (cliScope === 'disabled') {
-      return err(req.id, 'forbidden', 'CLI access is disabled for this agent group.');
-    }
-
     if (cliScope === 'group') {
-      // Only allow whitelisted resources and general commands (no resource, like help)
-      if (cmd.resource && !GROUP_SCOPE_RESOURCES.has(cmd.resource)) {
-        return err(req.id, 'forbidden', `CLI access is scoped to this agent group. Cannot access "${cmd.resource}".`);
-      }
-
-      // Enforce group scope on all agent-group-related args.
-      // Different resources use different arg names for the agent group ID.
-      // Only check --id for resources where it IS the agent group ID.
-      const groupArgs = ['agent_group_id', 'group'] as const;
-      for (const key of groupArgs) {
-        if (req.args[key] && req.args[key] !== ctx.agentGroupId) {
-          return err(req.id, 'forbidden', 'CLI access is scoped to this agent group.');
-        }
-      }
-      if (
-        (cmd.resource === 'groups' || cmd.resource === 'destinations') &&
-        req.args.id &&
-        req.args.id !== ctx.agentGroupId
-      ) {
-        return err(req.id, 'forbidden', 'CLI access is scoped to this agent group.');
-      }
-
-      // Block cli_scope changes from group-scoped agents (privilege escalation)
-      if (req.args.cli_scope !== undefined || req.args['cli-scope'] !== undefined) {
-        return err(req.id, 'forbidden', 'Cannot change cli_scope from a group-scoped agent.');
-      }
-
       // Auto-fill agent-group-related args so the agent doesn't need
       // to pass its own group ID explicitly.
       const fill: Record<string, unknown> = {
@@ -116,9 +100,19 @@ export async function dispatch(
     }
   }
 
+  const decision = guard(commandGuard(cmd.name), {
+    actor: actorFor(ctx),
+    payload: req.args,
+    grant: opts.grant ?? null,
+  });
+
+  if (decision.effect === 'deny') {
+    return err(req.id, 'forbidden', decision.reason);
+  }
+
   // `--help` interception: answer with the command's generated help instead of
-  // executing. Placed after scope enforcement (a group-scoped agent can't probe
-  // forbidden resources) and BEFORE approval gating — asking for help on an
+  // executing. Placed after the guard's deny (a group-scoped agent can't probe
+  // forbidden resources) and BEFORE hold execution — asking for help on an
   // approval-gated verb must never mint an approval card.
   if (req.args.help === true) {
     // Carry the help text in `human` too, so both clients print it verbatim
@@ -127,7 +121,12 @@ export async function dispatch(
     return { id: req.id, ok: true, data: helpText, human: helpText };
   }
 
-  if (ctx.caller !== 'host' && cmd.access === 'approval' && !opts.approved) {
+  if (decision.effect === 'hold') {
+    if (ctx.caller !== 'agent') {
+      // Holds only arise for agent callers; anything else is a guard bug —
+      // fail closed rather than card a ghost.
+      return err(req.id, 'forbidden', decision.reason);
+    }
     const session = getSession(ctx.sessionId);
     if (!session) {
       return err(req.id, 'handler-error', 'Session not found.');
@@ -215,10 +214,10 @@ export async function dispatch(
   }
 }
 
-registerApprovalHandler('cli_command', async ({ payload, notify }) => {
+registerApprovalHandler('cli_command', async ({ payload, approval, notify }) => {
   const frame = payload.frame as RequestFrame;
   const callerContext = parseCallerContext(payload.callerContext) ?? { caller: 'host' };
-  const response = await dispatch(frame, callerContext, { approved: true });
+  const response = await dispatch(frame, callerContext, { grant: approval });
 
   if (response.ok) {
     const data = typeof response.data === 'string' ? response.data : JSON.stringify(response.data, null, 2);

src/cli/guard.ts

@@ -0,0 +1,86 @@
+/**
+ * CLI guard adapter — the command registry's catalog derivation and
+ * structural baseline, moved verbatim out of dispatch.ts (guarded-actions
+ * phase 2). Declaration is registration: registry.register() derives one
+ * catalog entry per command from the CommandDef itself; no second file is
+ * edited when a command is added.
+ *
+ * The baseline carries today's decisions exactly:
+ *   host caller → allow (the 0600 socket is the auth story — in code,
+ *   unremovable by data);
+ *   cli_scope 'disabled' → deny; 'group' → resource allowlist, cross-group
+ *   arg denial, cli_scope-change denial;
+ *   access 'approval' for agent callers → hold for the group's admin chain.
+ *
+ * Arg auto-fill, the sessions-get existence oracle, and post-handler row
+ * filtering stay in dispatch.ts — mechanics, not policy.
+ */
+import { getContainerConfig } from '../db/container-configs.js';
+import { ALLOW, DENY, HOLD, type GuardedActionSpec, type GuardInput } from '../guard/index.js';
+import { GROUP_SCOPE_RESOURCES, type CommandDef } from './registry.js';
+
+/** Dotted catalog action name for a command. */
+export function commandGuardAction(cmd: Pick<CommandDef, 'name' | 'action'>): string {
+  return cmd.action ?? `cli.${cmd.name}`;
+}
+
+/** Catalog entry derived from a CommandDef at registration time. */
+export function commandGuardSpec(cmd: CommandDef): GuardedActionSpec {
+  return {
+    action: commandGuardAction(cmd),
+    approvalAction: cmd.access === 'approval' ? 'cli_command' : undefined,
+    // Bind a cli_command grant to the exact command it was approved for.
+    grantMatches: (grant) => {
+      try {
+        const payload = JSON.parse(grant.payload) as { frame?: { command?: string } };
+        return payload.frame?.command === cmd.name;
+      } catch {
+        return false;
+      }
+    },
+    baseline: (input) => commandBaseline(cmd, input),
+  };
+}
+
+function commandBaseline(cmd: CommandDef, input: GuardInput) {
+  const { actor } = input;
+  if (actor.kind === 'host') return ALLOW('host caller (trusted socket)');
+  if (actor.kind !== 'agent') return DENY('CLI commands accept host or agent callers only.');
+
+  const args = input.payload;
+  const cliScope = getContainerConfig(actor.agentGroupId)?.cli_scope ?? 'group';
+
+  if (cliScope === 'disabled') {
+    return DENY('CLI access is disabled for this agent group.');
+  }
+
+  if (cliScope === 'group') {
+    // Only allow whitelisted resources and general commands (no resource, like help)
+    if (cmd.resource && !GROUP_SCOPE_RESOURCES.has(cmd.resource)) {
+      return DENY(`CLI access is scoped to this agent group. Cannot access "${cmd.resource}".`);
+    }
+
+    // Enforce group scope on all agent-group-related args.
+    // Different resources use different arg names for the agent group ID.
+    // Only check --id for resources where it IS the agent group ID.
+    for (const key of ['agent_group_id', 'group'] as const) {
+      if (args[key] && args[key] !== actor.agentGroupId) {
+        return DENY('CLI access is scoped to this agent group.');
+      }
+    }
+    if ((cmd.resource === 'groups' || cmd.resource === 'destinations') && args.id && args.id !== actor.agentGroupId) {
+      return DENY('CLI access is scoped to this agent group.');
+    }
+
+    // Block cli_scope changes from group-scoped agents (privilege escalation)
+    if (args.cli_scope !== undefined || args['cli-scope'] !== undefined) {
+      return DENY('Cannot change cli_scope from a group-scoped agent.');
+    }
+  }
+
+  if (cmd.access === 'approval') {
+    return HOLD(`agent-initiated "${cmd.name}" requires admin approval`);
+  }
+
+  return ALLOW('open command');
+}

src/cli/registry.ts

@@ -8,6 +8,8 @@
  * registers the help commands, so the registry is populated before the host's
  * CLI server accepts connections.
  */
+import { defineGuardedAction, type GuardedAction } from '../guard/index.js';
+import { commandGuardSpec } from './guard.js';
 import type { CallerContext } from './frame.js';
 
 /**
@@ -23,6 +25,13 @@ export type CommandDef<TArgs = unknown, TData = unknown> = {
   name: string;
   description: string;
   access: Access;
+  /**
+   * Dotted guard-catalog action name (e.g. `roles.grant`,
+   * `groups.config.add-mcp-server`). Set by registerResource from the
+   * resource + verb; commands registered directly (help) fall back to
+   * `cli.<name>`.
+   */
+  action?: string;
   /**
    * The group-scope whitelist key. Under `cli_scope: 'group'` the dispatcher
    * only lets an agent run commands whose `resource` is on the whitelist
@@ -52,12 +61,26 @@ export type CommandDef<TArgs = unknown, TData = unknown> = {
 };
 
 const registry = new Map<string, CommandDef>();
+const commandGuards = new Map<string, GuardedAction>();
 
 export function register<TArgs, TData>(def: CommandDef<TArgs, TData>): void {
   if (registry.has(def.name)) {
     throw new Error(`CLI command "${def.name}" already registered`);
   }
   registry.set(def.name, def as CommandDef);
+  // Declaration is registration: every command gets a guard-catalog entry
+  // derived from its own definition, in the same call that registers it — a
+  // command cannot exist without a guard, and dispatch consults it by value.
+  commandGuards.set(def.name, defineGuardedAction(commandGuardSpec(def as CommandDef)));
+}
+
+/** The guard defined for a registered command — total for anything register() accepted. */
+export function commandGuard(name: string): GuardedAction {
+  const g = commandGuards.get(name);
+  if (!g) {
+    throw new Error(`CLI command "${name}" has no guard — was it registered through register()?`);
+  }
+  return g;
 }
 
 export function lookup(name: string): CommandDef | undefined {

src/delivery-actions.test.ts

@@ -4,7 +4,9 @@
  * `registerDeliveryAction` is the hook modules use to handle system-kind
  * outbound messages; `getDeliveryAction` is the read side that makes those
  * registrations behavior-testable. Goes red if either half of the registry
- * is removed or the two stop sharing the same map.
+ * is removed or the two stop sharing the same map. Every registration now
+ * carries a guard spec or an explicit unguarded(<reason>) declaration —
+ * omission is a type error.
  */
 import { describe, it, expect, vi } from 'vitest';
 
@@ -16,11 +18,14 @@ vi.mock('./container-runner.js', () => ({
 }));
 
 import { registerDeliveryAction, getDeliveryAction, type DeliveryActionHandler } from './delivery.js';
+import { defineGuardedAction, HOLD, unguarded } from './guard/index.js';
+
+const testUnguarded = unguarded('test — registry mechanics only');
 
 describe('delivery action registry', () => {
   it('getDeliveryAction returns the handler registerDeliveryAction registered', () => {
     const handler: DeliveryActionHandler = async () => {};
-    registerDeliveryAction('test_registry_action', handler);
+    registerDeliveryAction('test_registry_action', handler, testUnguarded);
     expect(getDeliveryAction('test_registry_action')).toBe(handler);
   });
 
@@ -31,8 +36,34 @@ describe('delivery action registry', () => {
   it('re-registering an action overwrites the previous handler', () => {
     const first: DeliveryActionHandler = async () => {};
     const second: DeliveryActionHandler = async () => {};
-    registerDeliveryAction('test_overwrite_action', first);
-    registerDeliveryAction('test_overwrite_action', second);
+    registerDeliveryAction('test_overwrite_action', first, testUnguarded);
+    registerDeliveryAction('test_overwrite_action', second, testUnguarded);
     expect(getDeliveryAction('test_overwrite_action')).toBe(second);
   });
+
+  it('refuses to replace a guard-wrapped action with an unguarded handler', () => {
+    const guardAction = defineGuardedAction({
+      action: 'test.guarded-overwrite',
+      baseline: () => HOLD('t'),
+    });
+    registerDeliveryAction('test_guarded_overwrite', async () => {}, {
+      guardAction,
+      requestHold: async () => {},
+    });
+
+    // Disarming the guard by re-registering unguarded must throw — otherwise
+    // the action's catalog entry would still exist while the live path runs
+    // unguarded.
+    expect(() => registerDeliveryAction('test_guarded_overwrite', async () => {}, testUnguarded)).toThrow(
+      /disarm the guard/,
+    );
+
+    // Re-registering WITH a spec stays allowed (a legitimate replacement
+    // keeps the action guarded).
+    registerDeliveryAction('test_guarded_overwrite', async () => {}, {
+      guardAction,
+      requestHold: async () => {},
+    });
+    expect(getDeliveryAction('test_guarded_overwrite')).toBeDefined();
+  });
 });

src/delivery.ts

@@ -20,12 +20,13 @@ import {
   markDeliveryFailed,
   migrateDeliveredTable,
 } from './db/session-db.js';
+import { guard, type GuardedAction, type Unguarded } from './guard/index.js';
 import { log } from './log.js';
 import { normalizeOptions } from './channels/ask-question.js';
 import { clearOutbox, openInboundDb, openOutboundDb, readOutboxFiles } from './session-manager.js';
 import { pauseTypingRefreshAfterDelivery, setTypingAdapter } from './modules/typing/index.js';
 import type { OutboundFile } from './channels/adapter.js';
-import type { Session } from './types.js';
+import type { PendingApproval, Session } from './types.js';
 
 const ACTIVE_POLL_MS = 1000;
 const SWEEP_POLL_MS = 60_000;
@@ -393,14 +394,19 @@ async function deliverMessage(
  * Delivery action registry.
  *
  * Modules register handlers for system-kind outbound message actions via
- * `registerDeliveryAction`. Core checks the registry first in
- * `handleSystemAction` and falls through to the inline switch when no
- * handler is registered. The switch will shrink as modules are extracted
- * (scheduling, approvals, agent-to-agent) and eventually only its default
- * branch remains.
+ * `registerDeliveryAction`. Unknown actions log "Unknown system action".
  *
- * Default when no handler registered and the switch doesn't match: log
- * "Unknown system action" and return.
+ * Privileged delivery actions (create_agent, install_packages,
+ * add_mcp_server) register with a guard spec: every path to the handler body
+ * — dispatch, approved replay, test lookup — goes through the guard consult
+ * (allow / hold / deny), so there is no unguarded route to it. On approve,
+ * the continuation re-enters the same entry carrying the approval row as its
+ * grant (`reenterGuardedDeliveryAction`), so the structural baseline is
+ * re-checked live. Plain actions (scheduling self-actions, the cli_request
+ * bridge — its inner commands are guarded at dispatch) register with an
+ * explicit `unguarded(<reason>)` declaration instead of a spec — omission is
+ * not representable, so the decision to run unguarded is visible, and
+ * justified, at the registration site.
  */
 export type DeliveryActionHandler = (
   content: Record<string, unknown>,
@@ -408,18 +414,115 @@ export type DeliveryActionHandler = (
   inDb: Database.Database,
 ) => Promise<void>;
 
-const actionHandlers = new Map<string, DeliveryActionHandler>();
+/** Handler shape for guard-wrapped actions — must not touch inDb (replays run without one). */
+export type GuardedDeliveryHandler = (content: Record<string, unknown>, session: Session) => Promise<void>;
 
-export function registerDeliveryAction(action: string, handler: DeliveryActionHandler): void {
-  if (actionHandlers.has(action)) {
-    log.warn('Delivery action handler overwritten', { action });
-  }
-  actionHandlers.set(action, handler);
+export interface DeliveryGuardSpec {
+  /** Guard action consulted before the handler runs — the defined value, not a name. */
+  guardAction: GuardedAction;
+  /**
+   * Domain validation that runs before the guard — malformed requests are
+   * answered (notify) without ever creating a hold. Return false to stop.
+   */
+  precheck?: (content: Record<string, unknown>, session: Session) => boolean | Promise<boolean>;
+  /** Create the hold (the domain's requestApproval call — card text lives with the domain). */
+  requestHold: (content: Record<string, unknown>, session: Session) => Promise<void>;
+  /** Tell the requester about a deny. */
+  onDeny?: (content: Record<string, unknown>, session: Session, reason: string) => void;
 }
 
-/** Look up a registered delivery-action handler. Lets module registrations be behavior-tested. */
+type DeliveryEntry =
+  | { guard: Unguarded; handler: DeliveryActionHandler }
+  | { guard: DeliveryGuardSpec; handler: GuardedDeliveryHandler };
+
+const deliveryActions = new Map<string, DeliveryEntry>();
+
+function isUnguardedEntry(entry: DeliveryEntry): entry is Extract<DeliveryEntry, { guard: Unguarded }> {
+  return 'reason' in entry.guard;
+}
+
+export function registerDeliveryAction(action: string, handler: DeliveryActionHandler, unguardedDecl: Unguarded): void;
+export function registerDeliveryAction(action: string, handler: GuardedDeliveryHandler, spec: DeliveryGuardSpec): void;
+export function registerDeliveryAction(
+  action: string,
+  handler: DeliveryActionHandler | GuardedDeliveryHandler,
+  guardDecl: DeliveryGuardSpec | Unguarded,
+): void {
+  const existing = deliveryActions.get(action);
+  if (existing) {
+    // Replacing a guard-wrapped action with an unguarded handler would
+    // disarm the guard while its catalog entry still exists — refuse. A
+    // skill that wants to extend a guarded action must compose at the
+    // module's exported functions instead, or re-register with a guard spec
+    // of its own.
+    if ('reason' in guardDecl && !isUnguardedEntry(existing)) {
+      throw new Error(
+        `delivery action "${action}" is guard-wrapped; re-registering it without a guard spec would disarm the guard`,
+      );
+    }
+    log.warn('Delivery action handler overwritten', { action });
+  }
+  // The overloads pair each handler shape with its declaration; the merged
+  // implementation signature erases that pairing, hence the one cast.
+  deliveryActions.set(action, { guard: guardDecl, handler } as DeliveryEntry);
+}
+
+async function runGuarded(
+  action: string,
+  entry: Extract<DeliveryEntry, { guard: DeliveryGuardSpec }>,
+  content: Record<string, unknown>,
+  session: Session,
+  grant: PendingApproval | null,
+): Promise<void> {
+  const spec = entry.guard;
+  if (spec.precheck && !(await spec.precheck(content, session))) return;
+
+  const decision = guard(spec.guardAction, {
+    actor: { kind: 'agent', agentGroupId: session.agent_group_id, sessionId: session.id },
+    payload: content,
+    grant,
+  });
+
+  if (decision.effect === 'deny') {
+    log.warn('Delivery action denied by guard', { action, reason: decision.reason });
+    spec.onDeny?.(content, session, decision.reason);
+    return;
+  }
+  if (decision.effect === 'hold') {
+    await spec.requestHold(content, session);
+    return;
+  }
+  await entry.handler(content, session);
+}
+
+/**
+ * Approve continuation for a guard-wrapped delivery action: re-enter the
+ * entry with the approval row as the grant. The guard treats the grant as
+ * hold-satisfied but re-runs the structural baseline, so approve-then-revoke
+ * does not execute. Domains register this as their approval handler in the
+ * same line that registers the action.
+ */
+export function reenterGuardedDeliveryAction(action: string) {
+  return async (ctx: { session: Session; payload: Record<string, unknown>; approval: PendingApproval }) => {
+    const entry = deliveryActions.get(action);
+    if (!entry || isUnguardedEntry(entry)) {
+      log.warn('Approved replay for an action that is not guard-wrapped — dropping', { action });
+      return;
+    }
+    await runGuarded(action, entry, ctx.payload, ctx.session, ctx.approval);
+  };
+}
+
+/**
+ * The invocable for a registered action — the raw handler for unguarded
+ * entries, the guard-consulting path for guarded ones. Dispatch and tests
+ * both come through here; there is no route around the guard.
+ */
 export function getDeliveryAction(action: string): DeliveryActionHandler | undefined {
-  return actionHandlers.get(action);
+  const entry = deliveryActions.get(action);
+  if (!entry) return undefined;
+  if (isUnguardedEntry(entry)) return entry.handler;
+  return (content, session) => runGuarded(action, entry, content, session, null);
 }
 
 /**
@@ -435,7 +538,7 @@ async function handleSystemAction(
   const action = content.action as string;
   log.info('System action from agent', { sessionId: session.id, action });
 
-  const registered = actionHandlers.get(action);
+  const registered = getDeliveryAction(action);
   if (registered) {
     await registered(content, session, inDb);
     return;

src/guard-conformance.ts

@@ -0,0 +1,70 @@
+/**
+ * Boot-time guard sanity — the one cross-registry invariant left to check
+ * after all import-time registrations have run.
+ *
+ * The old registry walk is gone: everything it detected is now
+ * unconstructible at the API level.
+ *   - A consult site cannot name a missing catalog entry — guard() takes the
+ *     GuardedAction VALUE returned by defineGuardedAction, so a dropped
+ *     module-edge import or a typo'd action is a compile error (and a forged
+ *     value is denied at runtime), not a silent allow.
+ *   - A handler cannot register unguarded by omission — every registry
+ *     (delivery actions, response handlers, interceptors, CLI commands)
+ *     requires a guard spec or an explicit unguarded(<reason>) declaration
+ *     at the registration site.
+ *
+ * What remains is completeness ACROSS registries: a guarded action that
+ * holds via `approvalAction` needs a registered approval handler, or an
+ * approved card resolves into nothing — the hold has no continuation. That
+ * pairing only exists once every module has loaded (catalog entries and
+ * approval handlers register from different modules), so it stays a boot
+ * check with the fail-closed posture: the host refuses to start, surfacing
+ * the mis-composition at skill-install time instead of at the first
+ * approved card.
+ */
+import { listGuardedActions } from './guard/index.js';
+import { log } from './log.js';
+import { getApprovalHandler } from './modules/approvals/primitive.js';
+
+/** Holding actions with no approve continuation. Empty = conformant. */
+export function grantContinuationGaps(): string[] {
+  return listGuardedActions()
+    .filter((spec) => spec.approvalAction && !getApprovalHandler(spec.approvalAction))
+    .map(
+      (spec) =>
+        `guarded action "${spec.action}" holds via approval action "${spec.approvalAction}" ` +
+        'but no approval handler is registered — an approved hold would have no continuation',
+    );
+}
+
+/**
+ * Boot check: refuse to start when a holding action has no continuation.
+ * Call after all import-time registrations (any point in main()).
+ */
+export function enforceGuardConformance(): void {
+  const gaps = grantContinuationGaps();
+  if (gaps.length === 0) return;
+
+  console.error(
+    [
+      '',
+      '='.repeat(64),
+      'NanoClaw stopped: guard conformance failure',
+      '='.repeat(64),
+      'A guarded action can hold for approval, but no approval handler is',
+      'registered for its approval action — an admin could click Approve',
+      'and nothing would execute. This usually means a module (or skill)',
+      'defined a holding baseline without registering its continuation.',
+      '',
+      ...gaps.map((g) => `  - ${g}`),
+      '',
+      'Register the approval handler (registerApprovalHandler) in the same',
+      'module that defines the guarded action, or drop approvalAction from',
+      'the definition if the action can never hold.',
+      '='.repeat(64),
+      '',
+    ].join('\n'),
+  );
+  log.error('Guard conformance failure — refusing to start', { gaps });
+  process.exit(1);
+}

src/guard/conformance.test.ts

@@ -0,0 +1,84 @@
+/**
+ * Guard conformance — the boot invariant, checked with the real registries.
+ *
+ * The old registry walk is gone: an unmapped consult or an undeclared
+ * unguarded registration is now unconstructible — guard() takes the defined
+ * GuardedAction value (a dropped module-edge import or typo'd name is a
+ * compile error), and every registry requires a guard spec or an explicit
+ * unguarded(<reason>) declaration. What's left to verify structurally is the
+ * cross-registry pairing the compiler can't see: every holding action has a
+ * registered approve continuation. The check runs here in CI and at every
+ * boot (enforceGuardConformance refuses to start) — CI can't see
+ * skill-installed registrations, the boot check can.
+ */
+import { describe, expect, it } from 'vitest';
+
+// Production barrels — side-effect imports populate the real registries.
+import '../cli/commands/index.js';
+import '../modules/index.js';
+import '../cli/delivery-action.js';
+import '../cli/dispatch.js'; // registers the cli_command approval handler
+
+import { commandGuard, listCommands } from '../cli/registry.js';
+import { grantContinuationGaps } from '../guard-conformance.js';
+import { getApprovalHandler } from '../modules/approvals/primitive.js';
+import { defineGuardedAction, listGuardedActions } from './guard-actions.js';
+import { HOLD } from './types.js';
+
+describe('guard conformance', () => {
+  it('the grant-continuation check (shared with the boot check) reports zero gaps', () => {
+    expect(grantContinuationGaps()).toEqual([]);
+  });
+
+  it('every holding action pairs with a registered approval handler', () => {
+    const holding = listGuardedActions().filter((spec) => spec.approvalAction);
+    expect(holding.length).toBeGreaterThan(0);
+
+    const dangling = holding.filter((spec) => !getApprovalHandler(spec.approvalAction as string));
+    expect(dangling.map((s) => s.action)).toEqual([]);
+  });
+
+  it('every mutating ncl command derives a guard that holds via cli_command', () => {
+    const mutating = listCommands().filter((cmd) => cmd.access === 'approval');
+    expect(mutating.length).toBeGreaterThan(0);
+
+    const wrong = mutating.filter((cmd) => commandGuard(cmd.name).approvalAction !== 'cli_command');
+    expect(wrong.map((c) => c.name)).toEqual([]);
+  });
+
+  it('the domain catalog entries are defined once the module barrels load', () => {
+    const actions = new Set(listGuardedActions().map((s) => s.action));
+    for (const expected of [
+      'agents.create',
+      'a2a.send',
+      'self_mod.install_packages',
+      'self_mod.add_mcp_server',
+      'senders.admit',
+      'channels.register',
+    ]) {
+      expect(actions.has(expected), `catalog is missing "${expected}"`).toBe(true);
+    }
+  });
+
+  it('defining the same action twice throws — names are the catalog key', () => {
+    defineGuardedAction({ action: 'test.dup-define', baseline: () => HOLD('x') });
+    expect(() => defineGuardedAction({ action: 'test.dup-define', baseline: () => HOLD('x') })).toThrow(
+      /already defined/,
+    );
+  });
+
+  // KEEP LAST: defines a holding action with no continuation into the shared
+  // per-worker catalog, so every gap check after this point sees it.
+  it('the check names a holding action with no approve continuation (what boot refuses on)', () => {
+    defineGuardedAction({
+      action: 'test.dangling-hold',
+      approvalAction: 'test_dangling_hold_approved',
+      baseline: () => HOLD('always'),
+    });
+
+    const gaps = grantContinuationGaps();
+    expect(gaps).toHaveLength(1);
+    expect(gaps[0]).toContain('test.dangling-hold');
+    expect(gaps[0]).toContain('no approval handler');
+  });
+});

src/guard/guard-actions.ts

@@ -0,0 +1,73 @@
+/**
+ * The action catalog — the enforcement boundary.
+ *
+ * An action either is defined here (and every consult passes its decision)
+ * or cannot be consulted at all: guard() takes the GuardedAction VALUE
+ * returned by defineGuardedAction, so the wiring between a consult site and
+ * its baseline is a symbol reference the compiler checks. A dropped
+ * module-edge import or a typo'd action name is a build error, not a
+ * runtime fail-open — there is no lookup that can miss.
+ *
+ * Definitions are still recorded by name so boot can enumerate them: the
+ * grant-continuation check (src/guard-conformance.ts) pairs every holding
+ * action with its registered approval handler, and duplicate names are
+ * refused at definition time (grants match on the name).
+ */
+import type { GuardDecision, GuardInput } from './types.js';
+import type { PendingApproval } from '../types.js';
+
+export interface GuardedActionSpec {
+  /** Dotted action name, e.g. 'roles.grant', 'agents.create', 'a2a.send'. */
+  action: string;
+  /**
+   * Today's structural checks for this action, verbatim — the only source of
+   * allow. Runs on every consult, including approved replays (a grant
+   * satisfies a hold, never a deny).
+   */
+  baseline: (input: GuardInput) => GuardDecision;
+  /**
+   * The pending_approvals.action its holds resolve through — a grant is only
+   * accepted when its row carries this action. Omit for actions that can
+   * never be held (deny/allow-only baselines).
+   */
+  approvalAction?: string;
+  /**
+   * Extra domain binding between a grant and the replayed input (e.g. the
+   * a2a target must match the held message). Runs in addition to the
+   * approvalAction + live-row checks.
+   */
+  grantMatches?: (grant: PendingApproval, input: GuardInput) => boolean;
+}
+
+declare const guardedActionBrand: unique symbol;
+/**
+ * A defined guarded action — only defineGuardedAction can mint one. The
+ * brand makes the type nominal: a hand-rolled { action, baseline } object
+ * does not typecheck at a consult site, and fails the runtime check too.
+ */
+export type GuardedAction = Readonly<GuardedActionSpec> & { readonly [guardedActionBrand]: true };
+
+const defined = new Map<string, GuardedAction>();
+const minted = new WeakSet<object>();
+
+export function defineGuardedAction(spec: GuardedActionSpec): GuardedAction {
+  if (defined.has(spec.action)) {
+    throw new Error(`guarded action "${spec.action}" is already defined — action names are the catalog key`);
+  }
+  const def = Object.freeze({ ...spec }) as GuardedAction;
+  minted.add(def);
+  defined.set(spec.action, def);
+  return def;
+}
+
+/**
+ * Runtime backstop for callers outside the type system (plain JS, casts):
+ * only values minted by defineGuardedAction pass — guard() denies the rest.
+ */
+export function isGuardedAction(value: unknown): value is GuardedAction {
+  return typeof value === 'object' && value !== null && minted.has(value);
+}
+
+export function listGuardedActions(): GuardedAction[] {
+  return [...defined.values()].sort((a, b) => a.action.localeCompare(b.action));
+}

src/guard/guard.test.ts

@@ -0,0 +1,160 @@
+/**
+ * Guard decision-function unit tests: the baseline is the decision (allow /
+ * hold / deny returned as-is), grant semantics (satisfies holds, never
+ * denies; invalid → refuse), the runtime backstop against forged action
+ * values, and the fail-closed posture on a throwing baseline.
+ *
+ * Uses synthetic actions defined per test — the catalog is per-worker module
+ * state with no reset, so action names are unique.
+ */
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+
+import { guard } from './guard.js';
+import { defineGuardedAction, type GuardedAction } from './guard-actions.js';
+import { ALLOW, DENY, HOLD, type GuardInput } from './types.js';
+
+const mockGetPendingApproval = vi.fn();
+vi.mock('../db/sessions.js', () => ({
+  getPendingApproval: (...args: unknown[]) => mockGetPendingApproval(...args),
+}));
+vi.mock('../log.js', () => ({
+  log: { info: vi.fn(), warn: vi.fn(), error: vi.fn(), debug: vi.fn() },
+}));
+
+const AGENT = { kind: 'agent', agentGroupId: 'ag-1', sessionId: 'sess-1' } as const;
+
+function input(extra: Partial<GuardInput> = {}): GuardInput {
+  return { actor: AGENT, payload: {}, ...extra };
+}
+
+beforeEach(() => {
+  mockGetPendingApproval.mockReset();
+});
+
+afterEach(() => {
+  vi.clearAllMocks();
+});
+
+describe('the baseline is the decision', () => {
+  it('baseline allow → allow', () => {
+    const action = defineGuardedAction({ action: 't.allow1', baseline: () => ALLOW('ok') });
+    expect(guard(action, input()).effect).toBe('allow');
+  });
+
+  it('baseline hold → hold, default approver chain', () => {
+    const action = defineGuardedAction({ action: 't.hold1', baseline: () => HOLD('needs approval') });
+    const d = guard(action, input());
+    expect(d.effect).toBe('hold');
+    if (d.effect === 'hold') {
+      expect(d.reason).toBe('needs approval');
+      expect(d.approverUserId).toBeUndefined();
+    }
+  });
+
+  it('baseline hold → hold, carrying a named approver', () => {
+    const action = defineGuardedAction({ action: 't.hold2', baseline: () => HOLD('policy row', 'telegram:dana') });
+    const d = guard(action, input());
+    expect(d.effect).toBe('hold');
+    if (d.effect === 'hold') expect(d.approverUserId).toBe('telegram:dana');
+  });
+
+  it('baseline deny → deny, carrying the reason', () => {
+    const action = defineGuardedAction({ action: 't.deny1', baseline: () => DENY('structurally unauthorized') });
+    const d = guard(action, input());
+    expect(d.effect).toBe('deny');
+    if (d.effect === 'deny') expect(d.reason).toBe('structurally unauthorized');
+  });
+
+  it('a forged action value (not from defineGuardedAction) is denied', () => {
+    const forged = { action: 't.forged', baseline: () => ALLOW('never vetted') } as unknown as GuardedAction;
+    const d = guard(forged, input());
+    expect(d.effect).toBe('deny');
+    if (d.effect === 'deny') expect(d.reason).toContain('undefined action');
+  });
+});
+
+describe('grants', () => {
+  const grantRow = (action: string) =>
+    ({ approval_id: 'appr-1', action, payload: '{}' }) as unknown as NonNullable<GuardInput['grant']>;
+
+  it('a valid live grant satisfies a hold', () => {
+    const action = defineGuardedAction({
+      action: 't.g1',
+      approvalAction: 'g1_approved',
+      baseline: () => HOLD('b'),
+    });
+    const grant = grantRow('g1_approved');
+    mockGetPendingApproval.mockReturnValue(grant);
+    expect(guard(action, input({ grant })).effect).toBe('allow');
+  });
+
+  it('a grant never satisfies a deny — the baseline is re-checked live', () => {
+    const action = defineGuardedAction({
+      action: 't.g2',
+      approvalAction: 'g2_approved',
+      baseline: () => DENY('revoked since'),
+    });
+    const grant = grantRow('g2_approved');
+    mockGetPendingApproval.mockReturnValue(grant);
+    const d = guard(action, input({ grant }));
+    expect(d.effect).toBe('deny');
+    if (d.effect === 'deny') expect(d.reason).toBe('revoked since');
+  });
+
+  it('a dead grant (row deleted) refuses instead of re-holding', () => {
+    const action = defineGuardedAction({
+      action: 't.g3',
+      approvalAction: 'g3_approved',
+      baseline: () => HOLD('b'),
+    });
+    mockGetPendingApproval.mockReturnValue(undefined);
+    const d = guard(action, input({ grant: grantRow('g3_approved') }));
+    expect(d.effect).toBe('deny');
+  });
+
+  it("a grant for a different action doesn't transfer", () => {
+    const action = defineGuardedAction({
+      action: 't.g4',
+      approvalAction: 'g4_approved',
+      baseline: () => HOLD('b'),
+    });
+    const grant = grantRow('other_action');
+    mockGetPendingApproval.mockReturnValue(grant);
+    expect(guard(action, input({ grant })).effect).toBe('deny');
+  });
+
+  it('a domain grantMatches binding can refuse a payload mismatch', () => {
+    const action = defineGuardedAction({
+      action: 't.g5',
+      approvalAction: 'g5_approved',
+      grantMatches: () => false,
+      baseline: () => HOLD('b'),
+    });
+    const grant = grantRow('g5_approved');
+    mockGetPendingApproval.mockReturnValue(grant);
+    expect(guard(action, input({ grant })).effect).toBe('deny');
+  });
+
+  it('a grant on an already-allowed action is a no-op', () => {
+    const action = defineGuardedAction({
+      action: 't.g6',
+      approvalAction: 'g6_approved',
+      baseline: () => ALLOW('ok'),
+    });
+    const grant = grantRow('g6_approved');
+    mockGetPendingApproval.mockReturnValue(grant);
+    expect(guard(action, input({ grant })).effect).toBe('allow');
+  });
+});
+
+describe('fail-closed posture', () => {
+  it('a throwing baseline denies', () => {
+    const action = defineGuardedAction({
+      action: 't.f1',
+      baseline: () => {
+        throw new Error('boom');
+      },
+    });
+    expect(guard(action, input()).effect).toBe('deny');
+  });
+});

src/guard/guard.ts

@@ -0,0 +1,72 @@
+/**
+ * guard() — the one decision function every privileged action consults.
+ *
+ * The decision is the action's structural baseline — today's code checks,
+ * defined per action at the module edges. The consult site holds the
+ * GuardedAction value itself (defineGuardedAction), so there is no name
+ * lookup and no fail-open path for an unknown action: an unwired consult is
+ * a compile error, and a value that didn't come from defineGuardedAction is
+ * denied at runtime. Policy-as-data (tighten-only rule sources composing
+ * with the baseline) is deliberately deferred to phase 3 of the
+ * guarded-actions design, where the generalized rules table arrives with its
+ * first operator-visible consumer; until then the one policy table
+ * (agent_message_policies) is consulted inside the a2a.send baseline.
+ *
+ * Grants: an approved replay carries the verified approval row. A valid
+ * grant (live pending row whose action matches the entry's approval action,
+ * plus any domain binding) satisfies a hold — the human already decided —
+ * but NEVER a deny: the baseline is re-checked live, so approve-then-revoke
+ * no longer executes. A grant that is present but invalid fails closed to
+ * deny (no second card).
+ *
+ * The guard itself fails closed: a throwing baseline denies.
+ */
+import { getPendingApproval } from '../db/sessions.js';
+import { log } from '../log.js';
+import { isGuardedAction, type GuardedAction } from './guard-actions.js';
+import { ALLOW, DENY, type GuardDecision, type GuardInput } from './types.js';
+
+export function guard(action: GuardedAction, input: GuardInput): GuardDecision {
+  if (!isGuardedAction(action)) {
+    // JS-level backstop — the branded type already forbids this. A
+    // hand-rolled object must not carry a baseline never vetted at
+    // definition time.
+    log.error('Guard consulted with an undefined action — failing closed', {
+      action: (action as { action?: unknown } | null)?.action,
+    });
+    return DENY('guard consulted with an undefined action (failing closed)');
+  }
+
+  let decision: GuardDecision;
+  try {
+    decision = action.baseline(input);
+  } catch (err) {
+    log.error('Guard evaluation threw — failing closed', { action: action.action, err });
+    return DENY('guard failure (failing closed)');
+  }
+
+  if (!input.grant || decision.effect !== 'hold') {
+    // A grant never loosens a deny (the baseline re-check is live), and a
+    // grant on an already-allowed action is a no-op.
+    return decision;
+  }
+
+  // An invalid grant on a replay is a refusal, not a fresh hold — approved
+  // replays must execute exactly once.
+  if (grantSatisfies(action, input)) {
+    return ALLOW(`hold satisfied by approval ${input.grant.approval_id}`);
+  }
+  return DENY('replay carried an invalid or mismatched grant');
+}
+
+function grantSatisfies(action: GuardedAction, input: GuardInput): boolean {
+  const grant = input.grant;
+  if (!grant || !action.approvalAction) return false;
+  if (grant.action !== action.approvalAction) return false;
+  // The row must still be live — resolution deletes it, so a grant can only
+  // execute once and a fabricated row object doesn't pass.
+  const live = getPendingApproval(grant.approval_id);
+  if (!live || live.action !== action.approvalAction) return false;
+  if (action.grantMatches && !action.grantMatches(grant, input)) return false;
+  return true;
+}

src/guard/index.ts

@@ -0,0 +1,28 @@
+/**
+ * Guard — the privileged-action decision seam (guarded-actions phase 2).
+ *
+ * See the guarded-actions decisions doc on the team hub. One decision
+ * function (guard.ts) and a definition-derived action catalog
+ * (guard-actions.ts). Consults carry the GuardedAction value returned by
+ * defineGuardedAction — never a name to look up — so mis-wiring is a build
+ * error, not a runtime fail-open.
+ * Domain-free leaf: domain baselines are defined at the domain modules' edges.
+ */
+export { guard } from './guard.js';
+export {
+  defineGuardedAction,
+  isGuardedAction,
+  listGuardedActions,
+  type GuardedAction,
+  type GuardedActionSpec,
+} from './guard-actions.js';
+export {
+  ALLOW,
+  DENY,
+  HOLD,
+  unguarded,
+  type GuardActor,
+  type GuardDecision,
+  type GuardInput,
+  type Unguarded,
+} from './types.js';

src/guard/types.ts

@@ -0,0 +1,64 @@
+/**
+ * Guard vocabulary — the decision seam every privileged action passes.
+ *
+ * The guard is a domain-free leaf: this module may import the DB read layer,
+ * config, log, and shared types — never src/cli/* or src/modules/*. Domain
+ * knowledge (what an action's structural baseline checks) arrives via
+ * definition: domain modules call defineGuardedAction (guard-actions.ts) at
+ * their module edges and pass the returned value to every consult and
+ * registration site — the wiring is a symbol reference the compiler checks.
+ */
+import type { PendingApproval } from '../types.js';
+
+/** Who is attempting the action. Mirrors the CLI CallerContext + click identities. */
+export type GuardActor =
+  | { kind: 'host' }
+  | { kind: 'agent'; agentGroupId: string; sessionId?: string }
+  | { kind: 'human'; userId: string }
+  | { kind: 'system' };
+
+export interface GuardInput {
+  actor: GuardActor;
+  /** Domain resource reference, e.g. { from, to } for a2a.send. */
+  resource?: Record<string, string>;
+  /** Action arguments — what the card summarizes and rules may later match on. */
+  payload: Record<string, unknown>;
+  /**
+   * Verified approval row carried by an approved replay. A valid grant
+   * satisfies a hold (the human already decided) but never a deny — the
+   * structural baseline is re-checked live on every replay.
+   */
+  grant?: PendingApproval | null;
+}
+
+declare const unguardedBrand: unique symbol;
+/**
+ * A registration that deliberately carries no guard. Omission is not
+ * representable — every registry requires either a guard spec or this
+ * marker, so the decision to run unguarded is visible, and justified, in
+ * the diff that registers the handler. The reason travels with the
+ * registration; `grep "unguarded("` is the complete inventory.
+ */
+export type Unguarded = { readonly reason: string; readonly [unguardedBrand]: true };
+
+export function unguarded(reason: string): Unguarded {
+  return Object.freeze({ reason }) as Unguarded;
+}
+
+export type GuardDecision =
+  | { effect: 'allow'; reason: string }
+  | { effect: 'hold'; reason: string; approverUserId?: string }
+  | { effect: 'deny'; reason: string };
+
+export const ALLOW = (reason: string): GuardDecision => ({ effect: 'allow', reason });
+export const DENY = (reason: string): GuardDecision => ({ effect: 'deny', reason });
+/**
+ * approverUserId names an exclusive approver for the hold (the a2a policy
+ * row's named approver). Absent, the hold goes to the approvals primitive's
+ * default chain (scoped admins → global admins → owners).
+ */
+export const HOLD = (reason: string, approverUserId?: string): GuardDecision => ({
+  effect: 'hold',
+  reason,
+  approverUserId,
+});

src/index.ts

@@ -14,6 +14,7 @@ import { initDb } from './db/connection.js';
 import { runMigrations } from './db/migrations/index.js';
 import { ensureContainerRuntimeRunning, cleanupOrphans } from './container-runtime.js';
 import { startActiveDeliveryPoll, startSweepDeliveryPoll, setDeliveryAdapter, stopDeliveryPolls } from './delivery.js';
+import { enforceGuardConformance } from './guard-conformance.js';
 import { startHostSweep, stopHostSweep } from './host-sweep.js';
 import { routeInbound } from './router.js';
 import { log } from './log.js';
@@ -69,6 +70,12 @@ async function main(): Promise<void> {
   // outside the sanctioned path (raw `git pull` instead of /update-nanoclaw).
   enforceUpgradeTripwire();
 
+  // 0.6 Guard conformance — every import-time registration has already run;
+  // refuse to start if any privileged command / delivery action is unmapped
+  // (CI can't see skill-installed code — this makes the invariant hold at
+  // the boundary where third-party registrations enter).
+  enforceGuardConformance();
+
   // 1. Init central DB
   const dbPath = path.join(DATA_DIR, 'v2.db');
   const db = initDb(dbPath);

src/modules/agent-to-agent/agent-route.ts

@@ -27,14 +27,15 @@ import { getAgentGroup } from '../../db/agent-groups.js';
 import { getInboundSourceSessionId, getMostRecentPeerSourceSessionId } from '../../db/session-db.js';
 import { getSession } from '../../db/sessions.js';
 import { wakeContainer } from '../../container-runner.js';
+import { guard } from '../../guard/index.js';
 import { log } from '../../log.js';
 import { openInboundDb, resolveSession, sessionDir, writeSessionMessage } from '../../session-manager.js';
-import type { Session } from '../../types.js';
+import type { PendingApproval, Session } from '../../types.js';
 import { requestApproval } from '../approvals/index.js';
-import { hasDestination } from './db/agent-destinations.js';
-import { getMessagePolicy } from './db/agent-message-policies.js';
+import { A2A_MESSAGE_GATE_ACTION, a2aSend } from './guard.js';
 
 export { isSafeAttachmentName };
+export { A2A_MESSAGE_GATE_ACTION } from './guard.js';
 
 export interface ForwardedAttachment {
   name: string;
@@ -230,56 +231,64 @@ function resolveTargetSession(msg: RoutableAgentMessage, sourceSession: Session,
   return resolveSession(targetAgentGroupId, null, null, 'agent-shared').session;
 }
 
-export async function routeAgentMessage(msg: RoutableAgentMessage, session: Session): Promise<void> {
+export async function routeAgentMessage(
+  msg: RoutableAgentMessage,
+  session: Session,
+  opts: { grant?: PendingApproval } = {},
+): Promise<void> {
   const sourceAgentGroupId = session.agent_group_id;
   const targetAgentGroupId = msg.platform_id;
   if (!targetAgentGroupId) {
     throw new Error(`agent-to-agent message ${msg.id} is missing a target agent group id`);
   }
-  const isSelf = targetAgentGroupId === sourceAgentGroupId;
-  if (!isSelf && !hasDestination(sourceAgentGroupId, 'agent', targetAgentGroupId)) {
-    throw new Error(`unauthorized agent-to-agent: ${sourceAgentGroupId} has no destination for ${targetAgentGroupId}`);
-  }
-  if (!getAgentGroup(targetAgentGroupId)) {
-    throw new Error(`target agent group ${targetAgentGroupId} not found for message ${msg.id}`);
+
+  // The a2a.send baseline (guard.ts) carries the checks verbatim in their
+  // original order: destination ACL deny, target-exists deny, self-send
+  // allow, agent_message_policies hold. An approved replay carries the
+  // grant — the hold is satisfied but the structure is re-checked live, so
+  // revoking a destination between hold and approve blocks delivery.
+  const decision = guard(a2aSend, {
+    actor: { kind: 'agent', agentGroupId: sourceAgentGroupId, sessionId: session.id },
+    resource: { from: sourceAgentGroupId, to: targetAgentGroupId },
+    payload: { id: msg.id, platform_id: targetAgentGroupId, content: msg.content, in_reply_to: msg.in_reply_to },
+    grant: opts.grant ?? null,
+  });
+
+  if (decision.effect === 'deny') {
+    throw new Error(decision.reason);
   }
 
   // Gated edge: hold the message and return (not throw) so the delivery loop
-  // consumes the outbound row; `applyA2aMessageGate` re-routes it on approve.
-  if (!isSelf) {
-    const policy = getMessagePolicy(sourceAgentGroupId, targetAgentGroupId);
-    if (policy) {
-      const { approver } = policy;
-      const sourceName = getAgentGroup(sourceAgentGroupId)?.name ?? sourceAgentGroupId;
-      const targetName = getAgentGroup(targetAgentGroupId)?.name ?? targetAgentGroupId;
-      await requestApproval({
-        session,
-        agentName: sourceName,
-        action: A2A_MESSAGE_GATE_ACTION,
-        approverUserId: approver,
-        title: 'Message approval',
-        question: buildGateQuestion(sourceName, targetName, msg.content),
-        payload: {
-          id: msg.id,
-          platform_id: targetAgentGroupId,
-          content: msg.content,
-          in_reply_to: msg.in_reply_to,
-        },
-      });
-      log.info('Agent message held for approval', {
-        from: sourceAgentGroupId,
-        to: targetAgentGroupId,
-        msgId: msg.id,
-      });
-      return;
-    }
+  // consumes the outbound row; `applyA2aMessageGate` re-enters here with the
+  // grant on approve.
+  if (decision.effect === 'hold') {
+    const sourceName = getAgentGroup(sourceAgentGroupId)?.name ?? sourceAgentGroupId;
+    const targetName = getAgentGroup(targetAgentGroupId)?.name ?? targetAgentGroupId;
+    await requestApproval({
+      session,
+      agentName: sourceName,
+      action: A2A_MESSAGE_GATE_ACTION,
+      approverUserId: decision.approverUserId,
+      title: 'Message approval',
+      question: buildGateQuestion(sourceName, targetName, msg.content),
+      payload: {
+        id: msg.id,
+        platform_id: targetAgentGroupId,
+        content: msg.content,
+        in_reply_to: msg.in_reply_to,
+      },
+    });
+    log.info('Agent message held for approval', {
+      from: sourceAgentGroupId,
+      to: targetAgentGroupId,
+      msgId: msg.id,
+    });
+    return;
   }
 
   await performAgentRoute(msg, session, targetAgentGroupId);
 }
 
-export const A2A_MESSAGE_GATE_ACTION = 'a2a_message_gate';
-
 const GATE_CARD_BODY_MAX = 1500;
 
 function parseMessageContent(contentStr: string): { text: string; files: string[] } {
@@ -308,9 +317,11 @@ function buildGateQuestion(sourceName: string, targetName: string, contentStr: s
 
 /**
  * Cross-session route: pick the target session, forward files, write to its
- * inbound DB, wake it. Authorization is the caller's responsibility.
+ * inbound DB, wake it. Module-private — the only door is routeAgentMessage's
+ * guard decision (the approve continuation re-enters with a grant rather
+ * than calling this directly).
  */
-export async function performAgentRoute(
+async function performAgentRoute(
   msg: RoutableAgentMessage,
   session: Session,
   targetAgentGroupId: string,

src/modules/agent-to-agent/create-agent.test.ts

@@ -2,26 +2,48 @@
  * Tests for create_agent host-side authorization.
  *
  * Regression guard for the audit finding: `create_agent` is a privileged
- * central-DB write with no host-side authz. The fix authorizes by CLI scope —
- * trusted owner agent groups ('global') create directly; confined groups
- * ('group', the default and the prompt-injection victim) must get admin
- * approval. These tests pin that branch decision.
+ * central-DB write with no host-side authz. Authorization is the guard's
+ * `agents.create` baseline — trusted owner agent groups ('global') create
+ * directly; confined groups ('group', the default and the prompt-injection
+ * victim) hold for admin approval. These tests drive the REAL wrapped
+ * delivery action (the only reachable path) and the approve continuation's
+ * grant-carrying re-entry.
  */
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
-import type { Session } from '../../types.js';
+import type { PendingApproval, Session } from '../../types.js';
 
 // Mocks for the collaborators the branch decides between / depends on.
-const mockRequestApproval = vi.fn().mockResolvedValue(undefined);
-const mockGetContainerConfig = vi.fn();
-const mockCreateAgentGroup = vi.fn();
-const mockInitGroupFilesystem = vi.fn();
-const mockUpdateScalars = vi.fn();
-const mockWriteDestinations = vi.fn();
-const mockNotifyWrite = vi.fn();
+// vi.hoisted: the module barrel import below runs before this file's const
+// initializers, and the mock factories close over this state.
+const {
+  mockRequestApproval,
+  mockGetContainerConfig,
+  mockCreateAgentGroup,
+  mockInitGroupFilesystem,
+  mockUpdateScalars,
+  mockWriteDestinations,
+  mockNotifyWrite,
+  liveApprovals,
+  approvalHandlers,
+} = vi.hoisted(() => ({
+  mockRequestApproval: vi.fn().mockResolvedValue(undefined),
+  mockGetContainerConfig: vi.fn(),
+  mockCreateAgentGroup: vi.fn(),
+  mockInitGroupFilesystem: vi.fn(),
+  mockUpdateScalars: vi.fn(),
+  mockWriteDestinations: vi.fn(),
+  mockNotifyWrite: vi.fn(),
+  liveApprovals: new Map<string, import('../../types.js').PendingApproval>(),
+  approvalHandlers: new Map<string, (ctx: Record<string, unknown>) => Promise<void>>(),
+}));
 
 vi.mock('../approvals/index.js', () => ({
   requestApproval: (...a: unknown[]) => mockRequestApproval(...a),
+  notifyAgent: vi.fn(),
+  registerApprovalHandler: (action: string, handler: (ctx: Record<string, unknown>) => Promise<void>) => {
+    approvalHandlers.set(action, handler);
+  },
 }));
 vi.mock('../../db/container-configs.js', () => ({
   getContainerConfig: (...a: unknown[]) => mockGetContainerConfig(...a),
@@ -42,36 +64,81 @@ vi.mock('./write-destinations.js', () => ({
 vi.mock('./db/agent-destinations.js', () => ({
   getDestinationByName: () => undefined,
   createDestination: vi.fn(),
+  hasDestination: () => true,
   normalizeName: (s: string) => s.toLowerCase().replace(/[^a-z0-9]+/g, '-'),
 }));
 // notifyAgent writes to the session inbound.db + wakes the container; stub both.
+// delivery.ts and agent-route.ts pull more session-manager exports at import time.
 vi.mock('../../session-manager.js', () => ({
   writeSessionMessage: (...a: unknown[]) => mockNotifyWrite(...a),
+  openInboundDb: vi.fn(),
+  openOutboundDb: vi.fn(),
+  clearOutbox: vi.fn(),
+  readOutboxFiles: vi.fn().mockReturnValue([]),
+  resolveSession: vi.fn(),
+  sessionDir: vi.fn().mockReturnValue('/tmp/nowhere'),
+  inboundDbPath: vi.fn().mockReturnValue('/tmp/nowhere/inbound.db'),
 }));
 vi.mock('../../container-runner.js', () => ({
   wakeContainer: vi.fn().mockResolvedValue(undefined),
 }));
 vi.mock('../../db/sessions.js', () => ({
   getSession: (id: string) => ({ id, agent_group_id: 'ag-1' }),
+  getPendingApproval: (id: string) => liveApprovals.get(id),
+  getRunningSessions: () => [],
+  getActiveSessions: () => [],
+  createPendingQuestion: vi.fn(),
 }));
 
-import { handleCreateAgent } from './create-agent.js';
+// The a2a module barrel registers ./guard.js (catalog entries) and the
+// guard-wrapped create_agent delivery action — the path under test.
+import './index.js';
+import { getDeliveryAction } from '../../delivery.js';
 
 const SESSION = { id: 'sess-1', agent_group_id: 'ag-1' } as Session;
 
+async function runCreateAgent(content: Record<string, unknown>): Promise<void> {
+  const wrapped = getDeliveryAction('create_agent');
+  expect(wrapped).toBeDefined();
+  await wrapped!(content, SESSION, undefined as never);
+}
+
+function liveGrant(approvalId: string, payload: Record<string, unknown>): PendingApproval {
+  const row = {
+    approval_id: approvalId,
+    session_id: SESSION.id,
+    request_id: approvalId,
+    action: 'create_agent',
+    payload: JSON.stringify(payload),
+    created_at: new Date().toISOString(),
+    agent_group_id: 'ag-1',
+    channel_type: null,
+    platform_id: null,
+    platform_message_id: null,
+    expires_at: null,
+    status: 'pending',
+    title: '',
+    options_json: '[]',
+    approver_user_id: null,
+  } as PendingApproval;
+  liveApprovals.set(approvalId, row);
+  return row;
+}
+
 beforeEach(() => {
   vi.clearAllMocks();
+  liveApprovals.clear();
 });
 
 afterEach(() => {
   vi.restoreAllMocks();
 });
 
-describe('handleCreateAgent — scope-based authorization', () => {
+describe('create_agent — guard-based authorization (wrapped delivery action)', () => {
   it('global scope: creates directly, no approval requested', async () => {
     mockGetContainerConfig.mockReturnValue({ cli_scope: 'global' });
 
-    await handleCreateAgent({ name: 'Scout', instructions: 'help' }, SESSION);
+    await runCreateAgent({ name: 'Scout', instructions: 'help' });
 
     expect(mockRequestApproval).not.toHaveBeenCalled();
     expect(mockCreateAgentGroup).toHaveBeenCalledTimes(1);
@@ -84,7 +151,7 @@ describe('handleCreateAgent — scope-based authorization', () => {
     // dropping the inheritance leaves the child provider-less (→ claude).
     mockGetContainerConfig.mockReturnValue({ cli_scope: 'global', provider: 'codex' });
 
-    await handleCreateAgent({ name: 'Scout', instructions: 'help' }, SESSION);
+    await runCreateAgent({ name: 'Scout', instructions: 'help' });
 
     expect(mockInitGroupFilesystem).toHaveBeenCalledWith(
       expect.anything(),
@@ -96,7 +163,7 @@ describe('handleCreateAgent — scope-based authorization', () => {
   it('claude creator leaves the child provider unset (built-in default)', async () => {
     mockGetContainerConfig.mockReturnValue({ cli_scope: 'global' }); // no provider
 
-    await handleCreateAgent({ name: 'Scout', instructions: 'help' }, SESSION);
+    await runCreateAgent({ name: 'Scout', instructions: 'help' });
 
     expect(mockUpdateScalars).not.toHaveBeenCalled();
   });
@@ -104,7 +171,7 @@ describe('handleCreateAgent — scope-based authorization', () => {
   it('group scope (default): requires approval, does NOT create directly', async () => {
     mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
 
-    await handleCreateAgent({ name: 'Scout', instructions: 'help' }, SESSION);
+    await runCreateAgent({ name: 'Scout', instructions: 'help' });
 
     expect(mockRequestApproval).toHaveBeenCalledTimes(1);
     expect(mockRequestApproval.mock.calls[0][0]).toMatchObject({ action: 'create_agent' });
@@ -115,7 +182,7 @@ describe('handleCreateAgent — scope-based authorization', () => {
   it('missing config: fails closed to approval (no direct create)', async () => {
     mockGetContainerConfig.mockReturnValue(undefined);
 
-    await handleCreateAgent({ name: 'Scout' }, SESSION);
+    await runCreateAgent({ name: 'Scout' });
 
     expect(mockRequestApproval).toHaveBeenCalledTimes(1);
     expect(mockCreateAgentGroup).not.toHaveBeenCalled();
@@ -124,7 +191,7 @@ describe('handleCreateAgent — scope-based authorization', () => {
   it('disabled/other scope: requires approval', async () => {
     mockGetContainerConfig.mockReturnValue({ cli_scope: 'disabled' });
 
-    await handleCreateAgent({ name: 'Scout' }, SESSION);
+    await runCreateAgent({ name: 'Scout' });
 
     expect(mockRequestApproval).toHaveBeenCalledTimes(1);
     expect(mockCreateAgentGroup).not.toHaveBeenCalled();
@@ -133,9 +200,58 @@ describe('handleCreateAgent — scope-based authorization', () => {
   it('empty name: neither creates nor requests approval', async () => {
     mockGetContainerConfig.mockReturnValue({ cli_scope: 'global' });
 
-    await handleCreateAgent({ name: '' }, SESSION);
+    await runCreateAgent({ name: '' });
 
     expect(mockRequestApproval).not.toHaveBeenCalled();
     expect(mockCreateAgentGroup).not.toHaveBeenCalled();
   });
 });
+
+describe('create_agent — approved replay (grant-carrying re-entry)', () => {
+  it('valid grant executes exactly once — baseline hold is satisfied, create runs', async () => {
+    mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+    const payload = { name: 'Scout', instructions: 'help' };
+    const approval = liveGrant('appr-ca-1', payload);
+
+    const continuation = approvalHandlers.get('create_agent');
+    expect(continuation).toBeDefined();
+    await continuation!({ session: SESSION, payload, approval, userId: 'telegram:admin', notify: vi.fn() });
+
+    expect(mockCreateAgentGroup).toHaveBeenCalledTimes(1);
+    expect(mockRequestApproval).not.toHaveBeenCalled(); // no second card
+  });
+
+  it('dead grant (row already resolved) refuses the replay', async () => {
+    mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+    const payload = { name: 'Scout', instructions: 'help' };
+    const approval = liveGrant('appr-ca-2', payload);
+    liveApprovals.delete('appr-ca-2'); // resolution consumed the row
+
+    await approvalHandlers.get('create_agent')!({
+      session: SESSION,
+      payload,
+      approval,
+      userId: 'telegram:admin',
+      notify: vi.fn(),
+    });
+
+    expect(mockCreateAgentGroup).not.toHaveBeenCalled();
+    expect(mockRequestApproval).not.toHaveBeenCalled(); // refused, not re-held
+  });
+
+  it('mismatched grant (approved for a different name) refuses the replay', async () => {
+    mockGetContainerConfig.mockReturnValue({ cli_scope: 'group' });
+    const approval = liveGrant('appr-ca-3', { name: 'OtherAgent' });
+
+    await approvalHandlers.get('create_agent')!({
+      session: SESSION,
+      payload: { name: 'Scout' },
+      approval,
+      userId: 'telegram:admin',
+      notify: vi.fn(),
+    });
+
+    expect(mockCreateAgentGroup).not.toHaveBeenCalled();
+    expect(mockRequestApproval).not.toHaveBeenCalled();
+  });
+});

src/modules/agent-to-agent/create-agent.ts

@@ -1,16 +1,18 @@
 /**
- * `create_agent` delivery-action handler.
+ * `create_agent` delivery-action bodies.
  *
  * SECURITY: `create_agent` writes to the CENTRAL DB (agent_groups,
  * container_configs, agent_destinations) and scaffolds host filesystem state —
  * a privileged operation a confined container is otherwise architecturally
  * barred from. The container's MCP tool gate is inside the (untrusted)
  * container and is trivially bypassed by writing the outbound system row
- * directly, so authorization MUST be enforced host-side. Trusted owner agent
- * groups (CLI scope 'global') create directly; every other (confined) group
- * requires admin approval via `requestApproval` — matching `ncl groups create`
- * (access: 'approval') and the self-mod actions. `applyCreateAgent` runs the
- * creation on approve; `performCreateAgent` is the shared body.
+ * directly, so authorization MUST be enforced host-side: the delivery
+ * registry wraps this action with the guard, whose `agents.create` baseline
+ * (./guard.ts) is the old cli_scope branch verbatim — trusted global-scope
+ * groups allow, everything else (including unknown config, fail-closed)
+ * holds for admin approval. On approve the continuation re-enters the
+ * wrapped action with the approval row as its grant and `createAgent` runs.
+ * `performCreateAgent` is the module-private body.
  */
 import path from 'path';
 
@@ -23,7 +25,7 @@ import { initGroupFilesystem } from '../../group-init.js';
 import { log } from '../../log.js';
 import { writeSessionMessage } from '../../session-manager.js';
 import type { AgentGroup, Session } from '../../types.js';
-import { requestApproval, type ApprovalHandler } from '../approvals/index.js';
+import { requestApproval } from '../approvals/index.js';
 import { createDestination, getDestinationByName, normalizeName } from './db/agent-destinations.js';
 import { writeDestinations } from './write-destinations.js';
 
@@ -43,41 +45,27 @@ function notifyAgent(session: Session, text: string): void {
   }
 }
 
-/**
- * Delivery-action entry.
- *
- * Authorization depends on the calling group's CLI scope:
- *   - `global` (set by init-first-agent for trusted owner agent groups):
- *     create immediately. create_agent is the intended primitive for these
- *     privileged agents, and an approval tap on every sub-agent spawn would be
- *     needless friction.
- *   - anything else (the default `group` scope — the realistic
- *     prompt-injection victim): require an admin to approve before any
- *     central-DB write. `applyCreateAgent` runs on approve.
- * Unknown/missing config fails closed to the approval path.
- */
-export async function handleCreateAgent(content: Record<string, unknown>, session: Session): Promise<void> {
+/** Guard precheck: malformed requests are answered without ever creating a hold. */
+export function validateCreateAgent(content: Record<string, unknown>, session: Session): boolean {
   const name = typeof content.name === 'string' ? content.name : '';
-  const instructions = typeof content.instructions === 'string' ? content.instructions : null;
-
   if (!name) {
     notifyAgent(session, 'create_agent failed: name is required.');
-    return;
+    return false;
   }
-
-  const sourceGroup = getAgentGroup(session.agent_group_id);
-  if (!sourceGroup) {
+  if (!getAgentGroup(session.agent_group_id)) {
     notifyAgent(session, 'create_agent failed: source agent group not found.');
     log.warn('create_agent failed: missing source group', { sessionAgentGroup: session.agent_group_id, name });
-    return;
+    return false;
   }
+  return true;
+}
 
-  const cliScope = getContainerConfig(session.agent_group_id)?.cli_scope ?? 'group';
-  if (cliScope === 'global') {
-    // Trusted owner agent group — create directly, then notify (+wake) it.
-    await performCreateAgent(name, instructions, session, sourceGroup, (text) => notifyAgent(session, text));
-    return;
-  }
+/** Guard hold: card the requesting group's admin chain. */
+export async function requestCreateAgentHold(content: Record<string, unknown>, session: Session): Promise<void> {
+  const name = typeof content.name === 'string' ? content.name : '';
+  const instructions = typeof content.instructions === 'string' ? content.instructions : null;
+  const sourceGroup = getAgentGroup(session.agent_group_id);
+  if (!sourceGroup) return;
 
   await requestApproval({
     session,
@@ -89,35 +77,22 @@ export async function handleCreateAgent(content: Record<string, unknown>, sessio
   });
 }
 
-/**
- * Approval handler: performs the creation once an admin approves a request from
- * a confined (non-global) agent group. `session` is the requesting parent.
- */
-export const applyCreateAgent: ApprovalHandler = async ({ session, payload, notify }) => {
-  const name = typeof payload.name === 'string' ? payload.name : '';
-  const instructions = typeof payload.instructions === 'string' ? payload.instructions : null;
-
-  if (!name) {
-    notify('create_agent approved but the request had no name.');
-    return;
-  }
-
+/** Guard allow body: performs the creation (fresh global-scope call or approved replay). */
+export async function createAgent(content: Record<string, unknown>, session: Session): Promise<void> {
+  const name = typeof content.name === 'string' ? content.name : '';
+  const instructions = typeof content.instructions === 'string' ? content.instructions : null;
   const sourceGroup = getAgentGroup(session.agent_group_id);
-  if (!sourceGroup) {
-    notify('create_agent approved but the source agent group no longer exists.');
-    log.warn('create_agent apply failed: missing source group', { sessionAgentGroup: session.agent_group_id, name });
-    return;
-  }
+  if (!name || !sourceGroup) return; // precheck already answered the requester
 
-  await performCreateAgent(name, instructions, session, sourceGroup, notify);
-};
+  await performCreateAgent(name, instructions, session, sourceGroup, (text) => notifyAgent(session, text));
+}
 
 /**
  * Core creation: writes the new agent group + bidirectional destinations and
  * scaffolds its filesystem, then reports via `notify`. Authorization is the
- * CALLER's responsibility (the global-scope shortcut in handleCreateAgent or
- * admin approval via applyCreateAgent) — never call this from an unauthorized
- * path, as it performs privileged central-DB writes a confined container is
+ * CALLER's responsibility (the guard's agents.create decision) — never call
+ * this from an unauthorized path, as it performs privileged central-DB
+ * writes a confined container is
  * otherwise barred from.
  */
 async function performCreateAgent(

src/modules/agent-to-agent/guard.ts

@@ -0,0 +1,88 @@
+/**
+ * Agent-to-agent guard adapter — the module's catalog entries, composed at
+ * the module edge (imported by ./index.ts).
+ *
+ * agents.create — the cli_scope branch moved verbatim out of
+ * create-agent.ts: `global` scope creates directly (create_agent is the
+ * intended primitive for trusted owner agent groups); anything else — the
+ * default `group` scope, and unknown/missing config, fail-closed — holds for
+ * the requesting group's admin chain.
+ *
+ * a2a.send — the decision moved verbatim out of routeAgentMessage, in its
+ * original check order: a missing destination row denies; a missing target
+ * group denies; self-sends allow without a destination row; an
+ * agent_message_policies row for the (from, to) pair holds for the row's
+ * named approver. The ghost-policy edge (policy row with no destination row)
+ * denies — the destination check precedes the policy check, exactly today's
+ * outcome. Policy rows can only tighten (hold), never allow: absence of a
+ * row falls through to the structural checks.
+ */
+import { getAgentGroup } from '../../db/agent-groups.js';
+import { getContainerConfig } from '../../db/container-configs.js';
+import { ALLOW, DENY, HOLD, defineGuardedAction } from '../../guard/index.js';
+import { hasDestination } from './db/agent-destinations.js';
+import { getMessagePolicy } from './db/agent-message-policies.js';
+
+/**
+ * pending_approvals action string for held a2a messages. Lives here (not in
+ * agent-route.ts) so agent-route can import this adapter — loading the
+ * consult site guarantees its catalog entry is registered — without a cycle.
+ */
+export const A2A_MESSAGE_GATE_ACTION = 'a2a_message_gate';
+
+export const agentsCreate = defineGuardedAction({
+  action: 'agents.create',
+  approvalAction: 'create_agent',
+  // Bind a create_agent grant to the name that was approved.
+  grantMatches: (grant, input) => {
+    try {
+      return (JSON.parse(grant.payload) as { name?: string }).name === input.payload.name;
+    } catch {
+      return false;
+    }
+  },
+  baseline: (input) => {
+    if (input.actor.kind !== 'agent') return DENY('create_agent is a container-originated action.');
+    const cliScope = getContainerConfig(input.actor.agentGroupId)?.cli_scope ?? 'group';
+    if (cliScope === 'global') {
+      // Trusted owner agent group — an approval tap on every sub-agent spawn
+      // would be needless friction.
+      return ALLOW('trusted global-scope agent group');
+    }
+    // The realistic prompt-injection victim (default `group` scope) — and any
+    // unknown config value, fail-closed — requires an admin before any
+    // central-DB write.
+    return HOLD('agent-initiated create_agent requires admin approval');
+  },
+});
+
+export const a2aSend = defineGuardedAction({
+  action: 'a2a.send',
+  approvalAction: A2A_MESSAGE_GATE_ACTION,
+  // Bind an a2a grant to the exact held message target.
+  grantMatches: (grant, input) => {
+    try {
+      return (JSON.parse(grant.payload) as { platform_id?: string }).platform_id === input.resource?.to;
+    } catch {
+      return false;
+    }
+  },
+  baseline: (input) => {
+    if (input.actor.kind !== 'agent') return DENY('agent-to-agent send requires an agent actor');
+    const from = input.actor.agentGroupId;
+    const to = input.resource?.to ?? '';
+    const isSelf = to === from;
+    if (!isSelf && !hasDestination(from, 'agent', to)) {
+      return DENY(`unauthorized agent-to-agent: ${from} has no destination for ${to}`);
+    }
+    if (!getAgentGroup(to)) {
+      return DENY(`target agent group ${to} not found for message ${String(input.payload.id)}`);
+    }
+    if (isSelf) return ALLOW('self-send');
+    const policy = getMessagePolicy(from, to);
+    if (policy) {
+      return HOLD(`a2a message policy ${from}→${to} holds for ${policy.approver}`, policy.approver);
+    }
+    return ALLOW('destination grant exists');
+  },
+});

src/modules/agent-to-agent/index.ts

@@ -1,13 +1,14 @@
 /**
  * Agent-to-agent module — inter-agent messaging and on-demand agent creation.
  *
- * Registers one delivery action (`create_agent`) plus its matching approval
- * handler — `create_agent` writes central-DB state, so confined (non-global)
- * groups require admin approval (the delivery action queues the request;
- * `applyCreateAgent` runs on approve); trusted global-scope groups create
- * directly. The sibling `channel_type === 'agent'` routing path is NOT a system
- * action — core `delivery.ts` dispatches into `./agent-route.js` via a dynamic
- * import when it sees `msg.channel_type === 'agent'`.
+ * Registers its guard-catalog entries (./guard.js) and one guard-wrapped
+ * delivery action (`create_agent`) — `create_agent` writes central-DB state,
+ * so the guard's agents.create baseline holds confined (non-global) groups
+ * for admin approval while trusted global-scope groups create directly; the
+ * approval handler re-enters the wrapped action carrying the approval row as
+ * its grant. The sibling `channel_type === 'agent'` routing path is NOT a
+ * system action — core `delivery.ts` dispatches into `./agent-route.js` via
+ * a dynamic import when it sees `msg.channel_type === 'agent'`.
  *
  * Host integration points:
  *   - `src/container-runner.ts::spawnContainer` dynamically imports
@@ -20,13 +21,19 @@
  * system action logs "Unknown system action", `channel_type='agent'` messages
  * throw because the module isn't installed.
  */
-import { registerDeliveryAction } from '../../delivery.js';
-import { registerApprovalHandler } from '../approvals/index.js';
+import { reenterGuardedDeliveryAction, registerDeliveryAction } from '../../delivery.js';
+import { notifyAgent, registerApprovalHandler } from '../approvals/index.js';
 import { A2A_MESSAGE_GATE_ACTION } from './agent-route.js';
-import { applyCreateAgent, handleCreateAgent } from './create-agent.js';
+import { createAgent, requestCreateAgentHold, validateCreateAgent } from './create-agent.js';
+import { agentsCreate } from './guard.js';
 import { applyA2aMessageGate } from './message-gate.js';
 
-registerDeliveryAction('create_agent', handleCreateAgent);
-registerApprovalHandler('create_agent', applyCreateAgent);
+registerDeliveryAction('create_agent', createAgent, {
+  guardAction: agentsCreate,
+  precheck: validateCreateAgent,
+  requestHold: requestCreateAgentHold,
+  onDeny: (_content, session, reason) => notifyAgent(session, `create_agent denied: ${reason}`),
+});
+registerApprovalHandler('create_agent', reenterGuardedDeliveryAction('create_agent'));
 
 registerApprovalHandler(A2A_MESSAGE_GATE_ACTION, applyA2aMessageGate);

src/modules/agent-to-agent/message-gate.test.ts

@@ -2,16 +2,17 @@ import Database from 'better-sqlite3';
 import fs from 'fs';
 import { describe, expect, it, beforeEach, afterEach, vi } from 'vitest';
 
+import './guard.js'; // register the a2a.send catalog entry (incl. the policy hold)
 import { routeAgentMessage } from './agent-route.js';
 import { createDestination, deleteDestination, deleteAllDestinationsTouching } from './db/agent-destinations.js';
 import { getMessagePolicy, removeMessagePolicy, setMessagePolicy } from './db/agent-message-policies.js';
 import { applyA2aMessageGate } from './message-gate.js';
 import { initTestDb, closeDb, runMigrations, createAgentGroup } from '../../db/index.js';
 import { getDb } from '../../db/connection.js';
-import { createSession } from '../../db/sessions.js';
+import { createPendingApproval, createSession, deletePendingApproval, getPendingApproval } from '../../db/sessions.js';
 import { requestApproval } from '../approvals/index.js';
 import { initSessionFolder, inboundDbPath } from '../../session-manager.js';
-import type { Session } from '../../types.js';
+import type { PendingApproval, Session } from '../../types.js';
 
 vi.mock('../../container-runner.js', () => ({
   wakeContainer: vi.fn().mockResolvedValue(undefined),
@@ -67,6 +68,23 @@ function makeSession(id: string, agentGroupId: string): Session {
   };
 }
 
+/** Seed a live a2a hold row (what requestApproval writes) and return it as the grant. */
+function seedA2aHold(approvalId: string, payload: Record<string, unknown>): PendingApproval {
+  createPendingApproval({
+    approval_id: approvalId,
+    session_id: 'sess-A',
+    request_id: approvalId,
+    action: 'a2a_message_gate',
+    payload: JSON.stringify(payload),
+    created_at: now(),
+    agent_group_id: A,
+    title: 'Message approval',
+    options_json: '[]',
+    approver_user_id: 'telegram:dana',
+  });
+  return getPendingApproval(approvalId)!;
+}
+
 describe('agent message policies', () => {
   let SA: Session;
   let SB: Session;
@@ -129,7 +147,7 @@ describe('agent message policies', () => {
     expect(requestApproval).not.toHaveBeenCalled();
   });
 
-  it('policy present → holds the message and requests approval from the policy approver scoped to the target', async () => {
+  it('policy present → holds the message and requests approval from the policy approver', async () => {
     setMessagePolicy(A, B, 'telegram:dana', now());
 
     await routeAgentMessage(
@@ -139,7 +157,7 @@ describe('agent message policies', () => {
 
     // Held: nothing routed to B.
     expect(readInbound(B, SB.id)).toHaveLength(0);
-    // One approval requested, to the policy's approver, scoped to the target group.
+    // One approval requested, to the policy's approver.
     expect(requestApproval).toHaveBeenCalledTimes(1);
     const opts = vi.mocked(requestApproval).mock.calls[0][0];
     expect(opts.action).toBe('a2a_message_gate');
@@ -158,21 +176,71 @@ describe('agent message policies', () => {
     expect(readInbound(A, SA.id)).toHaveLength(1);
   });
 
-  // ── approve handler re-routes the held message ──
+  it('ghost policy (policy row, no destination row) still denies — deny beats the policy hold', async () => {
+    deleteDestination(A, 'b'); // removes A→B — the destination ACL now denies
+    setMessagePolicy(A, B, 'telegram:dana', now()); // ...but a stale policy row remains
+
+    await expect(
+      routeAgentMessage({ id: 'ghost', platform_id: B, content: JSON.stringify({ text: 'x' }), in_reply_to: null }, SA),
+    ).rejects.toThrow(/unauthorized agent-to-agent/);
+    expect(requestApproval).not.toHaveBeenCalled();
+    expect(readInbound(B, SB.id)).toHaveLength(0);
+  });
+
+  // ── approve handler re-enters the guarded route with the grant ──
+
+  it('applyA2aMessageGate delivers the held message to the target (valid grant)', async () => {
+    setMessagePolicy(A, B, 'telegram:dana', now());
+    const payload = { id: 'held-1', platform_id: B, content: JSON.stringify({ text: 'approved!' }), in_reply_to: null };
+    const approval = seedA2aHold('appr-a2a-1', payload);
 
-  it('applyA2aMessageGate delivers the held message to the target', async () => {
     const notify = vi.fn();
-    await applyA2aMessageGate({
-      session: SA,
-      userId: 'slack:dana',
-      notify,
-      payload: { id: 'held-1', platform_id: B, content: JSON.stringify({ text: 'approved!' }), in_reply_to: null },
-    });
+    await applyA2aMessageGate({ session: SA, userId: 'telegram:dana', notify, payload, approval });
 
     const bRows = readInbound(B, SB.id);
     expect(bRows).toHaveLength(1);
     expect(JSON.parse(bRows[0].content).text).toBe('approved!');
     expect(notify).not.toHaveBeenCalled();
+    // The hold is satisfied by the grant — no second card.
+    expect(requestApproval).not.toHaveBeenCalled();
+  });
+
+  it('destination revoked between hold and approve → the approved replay is blocked', async () => {
+    setMessagePolicy(A, B, 'telegram:dana', now());
+    const payload = { id: 'held-2', platform_id: B, content: JSON.stringify({ text: 'stale' }), in_reply_to: null };
+    const approval = seedA2aHold('appr-a2a-2', payload);
+
+    deleteDestination(A, 'b'); // revoke A→B while the card is pending
+
+    await expect(
+      applyA2aMessageGate({ session: SA, userId: 'telegram:dana', notify: vi.fn(), payload, approval }),
+    ).rejects.toThrow(/unauthorized agent-to-agent/);
+    expect(readInbound(B, SB.id)).toHaveLength(0);
+  });
+
+  it('mismatched grant (held for another target) refuses the replay', async () => {
+    setMessagePolicy(A, B, 'telegram:dana', now());
+    // Grant was approved for a message to A (different target than the replay).
+    const approval = seedA2aHold('appr-a2a-3', { id: 'other', platform_id: A, content: '{}', in_reply_to: null });
+    const payload = { id: 'held-3', platform_id: B, content: JSON.stringify({ text: 'swap' }), in_reply_to: null };
+
+    await expect(
+      applyA2aMessageGate({ session: SA, userId: 'telegram:dana', notify: vi.fn(), payload, approval }),
+    ).rejects.toThrow(/invalid or mismatched grant/);
+    expect(readInbound(B, SB.id)).toHaveLength(0);
+  });
+
+  it('a grant only works while its row is live (executes once)', async () => {
+    setMessagePolicy(A, B, 'telegram:dana', now());
+    const payload = { id: 'held-4', platform_id: B, content: JSON.stringify({ text: 'once' }), in_reply_to: null };
+    const approval = seedA2aHold('appr-a2a-4', payload);
+
+    deletePendingApproval(approval.approval_id); // resolution already consumed the row
+
+    await expect(
+      applyA2aMessageGate({ session: SA, userId: 'telegram:dana', notify: vi.fn(), payload, approval }),
+    ).rejects.toThrow(/invalid or mismatched grant/);
+    expect(readInbound(B, SB.id)).toHaveLength(0);
   });
 
   // ── ghost-gate cleanup ──

src/modules/agent-to-agent/message-gate.ts

@@ -1,9 +1,9 @@
 /** Approve handler for a held a2a message. (Reject is handled by the generic response-handler path.) */
 import { log } from '../../log.js';
 import type { ApprovalHandler } from '../approvals/index.js';
-import { performAgentRoute, type RoutableAgentMessage } from './agent-route.js';
+import { routeAgentMessage, type RoutableAgentMessage } from './agent-route.js';
 
-export const applyA2aMessageGate: ApprovalHandler = async ({ session, payload, notify }) => {
+export const applyA2aMessageGate: ApprovalHandler = async ({ session, payload, approval, notify }) => {
   const { id, platform_id, content, in_reply_to } = payload;
   if (typeof platform_id !== 'string' || !platform_id) {
     notify('Message approved but the target agent group was missing from the request.');
@@ -18,7 +18,12 @@ export const applyA2aMessageGate: ApprovalHandler = async ({ session, payload, n
     in_reply_to: typeof in_reply_to === 'string' ? in_reply_to : null,
   };
 
-  await performAgentRoute(msg, session, platform_id);
+  // One replay semantics: re-enter the guarded route carrying the approval
+  // row as the grant. The policy hold is satisfied, but the structural
+  // baseline runs live — un-wiring the pair between hold and approve now
+  // blocks delivery (the throw surfaces via the response handler's
+  // "approved, but applying it failed" notify).
+  await routeAgentMessage(msg, session, { grant: approval });
   log.info('Held agent message delivered after approval', {
     from: session.agent_group_id,
     to: platform_id,

src/modules/approvals/index.ts

@@ -23,6 +23,7 @@
  * + approval handlers via this module's public API.
  */
 import { onDeliveryAdapterReady } from '../../delivery.js';
+import { unguarded } from '../../guard/index.js';
 import { registerResponseHandler, onShutdown } from '../../response-registry.js';
 import { handleApprovalsResponse } from './response-handler.js';
 import { startOneCLIApprovalHandler, stopOneCLIApprovalHandler } from './onecli-approvals.js';
@@ -34,7 +35,10 @@ export type { ApprovalHandler, ApprovalHandlerContext, RequestApprovalOptions }
 // loads reason-capture.js, registering its message-interceptor on import.
 export { sweepAwaitingReasonRejects } from './reason-capture.js';
 
-registerResponseHandler(handleApprovalsResponse);
+registerResponseHandler(
+  handleApprovalsResponse,
+  unguarded('self-authorizing — resolves each pending_approvals row against its own approver rules before dispatching'),
+);
 
 onDeliveryAdapterReady((adapter) => {
   startOneCLIApprovalHandler(adapter);

src/modules/approvals/primitive.ts

@@ -59,6 +59,12 @@ const APPROVAL_OPTIONS: RawOption[] = [
 export interface ApprovalHandlerContext {
   session: Session;
   payload: Record<string, unknown>;
+  /**
+   * The verified approval row — the grant an approved continuation carries
+   * when it re-enters its guarded entry point. Still live here; resolution
+   * deletes it after the handler returns, so a grant executes exactly once.
+   */
+  approval: PendingApproval;
   /** User ID of the admin who approved. Empty string if unknown. */
   userId: string;
   /** Send a system chat message to the requesting agent's session. */

src/modules/approvals/reason-capture.ts

@@ -20,6 +20,7 @@
  */
 import type { InboundEvent } from '../../channels/adapter.js';
 import { getDeliveryAdapter } from '../../delivery.js';
+import { unguarded } from '../../guard/index.js';
 import {
   deletePendingApproval,
   getExpiredAwaitingReasonApprovals,
@@ -151,7 +152,10 @@ export async function captureReasonReply(event: InboundEvent): Promise<boolean>
   return true;
 }
 
-registerMessageInterceptor(captureReasonReply);
+registerMessageInterceptor(
+  captureReasonReply,
+  unguarded('self-authorizing — only captures a DM from the admin whose reject click armed it (keyed by userId + DM)'),
+);
 
 /**
  * Host-sweep finalizer: any reject-with-reason hold whose window elapsed (admin

src/modules/approvals/response-handler.ts

@@ -112,7 +112,7 @@ async function handleRegisteredApproval(
 
   const payload = JSON.parse(approval.payload);
   try {
-    await handler({ session, payload, userId, notify });
+    await handler({ session, payload, approval, userId, notify });
     log.info('Approval handled', { approvalId: approval.approval_id, action: approval.action, userId });
   } catch (err) {
     log.error('Approval handler threw', { approvalId: approval.approval_id, action: approval.action, err });

src/modules/interactive/index.ts

@@ -13,6 +13,7 @@
 import { getDb, hasTable } from '../../db/connection.js';
 import { deletePendingQuestion, getPendingQuestion, getSession } from '../../db/sessions.js';
 import { wakeContainer } from '../../container-runner.js';
+import { unguarded } from '../../guard/index.js';
 import { registerResponseHandler, type ResponsePayload } from '../../response-registry.js';
 import { log } from '../../log.js';
 import { writeSessionMessage } from '../../session-manager.js';
@@ -56,4 +57,7 @@ async function handleInteractiveResponse(payload: ResponsePayload): Promise<bool
   return true;
 }
 
-registerResponseHandler(handleInteractiveResponse);
+registerResponseHandler(
+  handleInteractiveResponse,
+  unguarded('not privileged — relays an in-chat answer back into the asking session; only the question row is touched'),
+);

src/modules/permissions/channel-approval.test.ts

@@ -54,7 +54,11 @@ vi.mock('./user-dm.js', () => ({
 
 vi.mock('../../config.js', async () => {
   const actual = await vi.importActual('../../config.js');
-  return { ...actual, DATA_DIR: '/tmp/nanoclaw-test-channel-approval' };
+  return {
+    ...actual,
+    DATA_DIR: '/tmp/nanoclaw-test-channel-approval',
+    GROUPS_DIR: '/tmp/nanoclaw-test-channel-approval/groups',
+  };
 });
 
 const TEST_DIR = '/tmp/nanoclaw-test-channel-approval';
@@ -438,6 +442,108 @@ describe('unknown-channel registration flow', () => {
       .c;
     expect(stillPending).toBe(1);
   });
+
+  it('create new agent: the free-text name reply creates the group and wires the channel', async () => {
+    const { routeInbound } = await import('../../router.js');
+    const { getResponseHandlers } = await import('../../response-registry.js');
+    const { getDb } = await import('../../db/connection.js');
+
+    await routeInbound(groupMention('chat-create-new'));
+    await new Promise((r) => setTimeout(r, 10));
+    const pending = getDb().prepare('SELECT messaging_group_id FROM pending_channel_approvals').get() as {
+      messaging_group_id: string;
+    };
+
+    // Owner clicks "Connect new agent" → name prompt lands in their DM.
+    for (const handler of getResponseHandlers()) {
+      const claimed = await handler({
+        questionId: pending.messaging_group_id,
+        value: 'new_agent',
+        userId: 'owner',
+        channelType: 'telegram',
+        platformId: 'dm-owner',
+        threadId: null,
+      });
+      if (claimed) break;
+    }
+
+    // Owner replies with the agent name in the same DM — the guarded
+    // interceptor allows (still an eligible approver) and creates.
+    await routeInbound({
+      channelType: 'telegram',
+      platformId: 'dm-owner',
+      threadId: null,
+      message: {
+        id: 'name-reply-1',
+        kind: 'chat' as const,
+        content: JSON.stringify({ senderId: 'owner', senderName: 'Owner', text: 'Newbie' }),
+        timestamp: now(),
+      },
+    });
+
+    const created = getDb().prepare("SELECT id FROM agent_groups WHERE name = 'Newbie'").get() as
+      | { id: string }
+      | undefined;
+    expect(created).toBeDefined();
+    const mgaCount = (
+      getDb()
+        .prepare('SELECT COUNT(*) AS c FROM messaging_group_agents WHERE messaging_group_id = ? AND agent_group_id = ?')
+        .get(pending.messaging_group_id, created!.id) as { c: number }
+    ).c;
+    expect(mgaCount).toBe(1);
+    const stillPending = (getDb().prepare('SELECT COUNT(*) AS c FROM pending_channel_approvals').get() as { c: number })
+      .c;
+    expect(stillPending).toBe(0);
+  });
+
+  it('a name reply after the registration vanished is consumed without creating anything', async () => {
+    const { routeInbound } = await import('../../router.js');
+    const { getResponseHandlers } = await import('../../response-registry.js');
+    const { getDb } = await import('../../db/connection.js');
+
+    await routeInbound(groupMention('chat-vanished'));
+    await new Promise((r) => setTimeout(r, 10));
+    const pending = getDb().prepare('SELECT messaging_group_id FROM pending_channel_approvals').get() as {
+      messaging_group_id: string;
+    };
+
+    for (const handler of getResponseHandlers()) {
+      const claimed = await handler({
+        questionId: pending.messaging_group_id,
+        value: 'new_agent',
+        userId: 'owner',
+        channelType: 'telegram',
+        platformId: 'dm-owner',
+        threadId: null,
+      });
+      if (claimed) break;
+    }
+
+    // The registration disappears between the click and the reply (rejected
+    // from another card, group delete cascade, …) — the guard's baseline no
+    // longer finds a pending registration, so the reply must not create.
+    getDb()
+      .prepare('DELETE FROM pending_channel_approvals WHERE messaging_group_id = ?')
+      .run(pending.messaging_group_id);
+
+    const agentGroupsBefore = (getDb().prepare('SELECT COUNT(*) AS c FROM agent_groups').get() as { c: number }).c;
+    await routeInbound({
+      channelType: 'telegram',
+      platformId: 'dm-owner',
+      threadId: null,
+      message: {
+        id: 'name-reply-2',
+        kind: 'chat' as const,
+        content: JSON.stringify({ senderId: 'owner', senderName: 'Owner', text: 'Ghost' }),
+        timestamp: now(),
+      },
+    });
+
+    const agentGroupsAfter = (getDb().prepare('SELECT COUNT(*) AS c FROM agent_groups').get() as { c: number }).c;
+    expect(agentGroupsAfter).toBe(agentGroupsBefore);
+    const mgaCount = (getDb().prepare('SELECT COUNT(*) AS c FROM messaging_group_agents').get() as { c: number }).c;
+    expect(mgaCount).toBe(0);
+  });
 });
 
 describe('no-owner / no-agent failure modes', () => {

src/modules/permissions/guard.ts

@@ -0,0 +1,54 @@
+/**
+ * Permissions guard adapter — the module's catalog entries, composed at the
+ * module edge (imported by ./index.ts).
+ *
+ * senders.admit — the `unknown_sender_policy` switch moved verbatim out of
+ * handleUnknownSender: `public` allows (short-circuited before the gate
+ * anyway), `request_approval` holds, `strict` denies. The hold is executed by
+ * the caller through the module's own pending_sender_approvals flow (card,
+ * in-flight dedup) — not the approvals primitive — so this entry has no
+ * approvalAction: the approve continuation adds the member and replays
+ * routeInbound, which then passes the gate structurally via membership, no
+ * grant needed.
+ *
+ * channels.register — click/reply authorization for the channel-registration
+ * flow, verbatim from today's response handler: the delivered approver, or an
+ * admin of the pending row's anchor agent group. Consulted by the wrapped
+ * response handler (card clicks) and the wrapped name-capture interceptor
+ * (free-text replies), so a privilege revoked mid-flow is re-checked at each
+ * step.
+ */
+import { ALLOW, DENY, HOLD, defineGuardedAction } from '../../guard/index.js';
+import { getPendingChannelApproval } from './db/pending-channel-approvals.js';
+import { hasAdminPrivilege } from './db/user-roles.js';
+
+export const sendersAdmit = defineGuardedAction({
+  action: 'senders.admit',
+  baseline: (input) => {
+    const policy = input.payload.policy;
+    if (policy === 'public') return ALLOW('public messaging group');
+    if (policy === 'request_approval') {
+      return HOLD(
+        `unknown sender requires admin approval on messaging group ${String(input.payload.messagingGroupId)}`,
+      );
+    }
+    return DENY('unknown sender on a strict messaging group');
+  },
+});
+
+export const channelsRegister = defineGuardedAction({
+  action: 'channels.register',
+  baseline: (input) => {
+    if (input.actor.kind !== 'human') return DENY('channel registration resolves via human clicks/replies');
+    const questionId = typeof input.payload.questionId === 'string' ? input.payload.questionId : '';
+    const row = getPendingChannelApproval(questionId);
+    if (!row) return DENY(`no pending channel registration for ${questionId || '(missing questionId)'}`);
+    if (
+      input.actor.userId &&
+      (input.actor.userId === row.approver_user_id || hasAdminPrivilege(input.actor.userId, row.agent_group_id))
+    ) {
+      return ALLOW('delivered approver or anchor-group admin');
+    }
+    return DENY('not an eligible channel-registration approver');
+  },
+});

src/modules/permissions/index.ts

@@ -32,6 +32,8 @@ import { registerResponseHandler, type ResponsePayload } from '../../response-re
 import { getDeliveryAdapter } from '../../delivery.js';
 import { log } from '../../log.js';
 import type { MessagingGroup, MessagingGroupAgent } from '../../types.js';
+import { guard, unguarded } from '../../guard/index.js';
+import { channelsRegister, sendersAdmit } from './guard.js';
 import { canAccessAgentGroup } from './access.js';
 import {
   buildAgentSelectionOptions,
@@ -129,43 +131,49 @@ function handleUnknownSender(
     agent_group_id: agentGroupId,
   };
 
-  if (mg.unknown_sender_policy === 'strict') {
-    log.info('MESSAGE DROPPED — unknown sender (strict policy)', {
+  // The admission decision is the guard's senders.admit baseline (./guard.ts)
+  // — unknown_sender_policy verbatim: strict → deny, request_approval → hold,
+  // public → allow (short-circuited before the gate). Drop-recording and the
+  // hold creation stay here.
+  const decision = guard(sendersAdmit, {
+    actor: userId ? { kind: 'human', userId } : { kind: 'system' },
+    payload: {
+      messagingGroupId: mg.id,
+      agentGroupId,
+      senderIdentity: userId,
+      policy: mg.unknown_sender_policy,
+    },
+  });
+
+  if (decision.effect === 'allow') return; // 'public' — handled before the gate; fall through silently.
+
+  log.info(
+    decision.effect === 'hold'
+      ? 'MESSAGE DROPPED — unknown sender (approval requested)'
+      : 'MESSAGE DROPPED — unknown sender (strict policy)',
+    {
       messagingGroupId: mg.id,
       agentGroupId,
       userId,
       accessReason,
-    });
-    recordDroppedMessage(dropRecord);
-    return;
-  }
+    },
+  );
+  recordDroppedMessage(dropRecord);
 
-  if (mg.unknown_sender_policy === 'request_approval') {
-    log.info('MESSAGE DROPPED — unknown sender (approval requested)', {
+  // Fire-and-forget; pick-approver + delivery + row-insert are all async.
+  // If it fails it logs internally — the user's message still stays dropped
+  // either way. Requires a resolved userId (senderResolver populates users
+  // row before the gate fires); if we got here without one, there's nothing
+  // to identify for approval and we just drop silently.
+  if (decision.effect === 'hold' && userId) {
+    requestSenderApproval({
       messagingGroupId: mg.id,
       agentGroupId,
-      userId,
-      accessReason,
-    });
-    recordDroppedMessage(dropRecord);
-    // Fire-and-forget; pick-approver + delivery + row-insert are all async.
-    // If it fails it logs internally — the user's message still stays dropped
-    // either way. Requires a resolved userId (senderResolver populates users
-    // row before the gate fires); if we got here without one, there's nothing
-    // to identify for approval and we just stay in the "silent strict" branch.
-    if (userId) {
-      requestSenderApproval({
-        messagingGroupId: mg.id,
-        agentGroupId,
-        senderIdentity: userId,
-        senderName,
-        event,
-      }).catch((err) => log.error('Sender-approval flow threw', { err }));
-    }
-    return;
+      senderIdentity: userId,
+      senderName,
+      event,
+    }).catch((err) => log.error('Sender-approval flow threw', { err }));
   }
-
-  // 'public' should have been handled before the gate; fall through silently.
 }
 
 setSenderResolver(extractAndUpsertUser);
@@ -285,7 +293,12 @@ async function handleSenderApprovalResponse(payload: ResponsePayload): Promise<b
   return true;
 }
 
-registerResponseHandler(handleSenderApprovalResponse);
+registerResponseHandler(
+  handleSenderApprovalResponse,
+  unguarded(
+    'self-authorizing — verifies the clicker inline (delivered approver or group admin) before adding a member',
+  ),
+);
 
 // ── Unknown-channel registration flow ──
 
@@ -311,21 +324,14 @@ async function handleChannelApprovalResponse(payload: ResponsePayload): Promise<
   const row = getPendingChannelApproval(payload.questionId);
   if (!row) return false;
 
+  // Click authorization is the guard's channels.register baseline (./guard.ts),
+  // consulted by the response-registry wrapper before this handler runs.
   const clickerId = payload.userId
     ? payload.userId.includes(':')
       ? payload.userId
       : `${payload.channelType}:${payload.userId}`
     : null;
-  const isAuthorized =
-    clickerId !== null && (clickerId === row.approver_user_id || hasAdminPrivilege(clickerId, row.agent_group_id));
-  if (!isAuthorized) {
-    log.warn('Channel registration click rejected — unauthorized clicker', {
-      messagingGroupId: row.messaging_group_id,
-      clickerId,
-      expectedApprover: row.approver_user_id,
-    });
-    return true;
-  }
+  if (!clickerId) return true; // unreachable behind the guard wrapper; fail closed
   const approverId = clickerId;
 
   // ── Reject / Cancel ──
@@ -515,13 +521,19 @@ async function handleChannelApprovalResponse(payload: ResponsePayload): Promise<
   return true;
 }
 
-registerResponseHandler(handleChannelApprovalResponse);
+registerResponseHandler(handleChannelApprovalResponse, {
+  action: channelsRegister,
+  claims: (payload) => getPendingChannelApproval(payload.questionId) !== undefined,
+});
 
 // ── Free-text name interceptor ──
 // Captures the next DM from an approver who clicked "Create new agent",
-// creates the agent immediately, wires the channel, and replays.
+// creates the agent immediately, wires the channel, and replays. The router
+// wraps it with the guard: the free-texter must still be an eligible
+// channel-registration approver at reply time — a privilege revoked between
+// the click and the reply now denies, and the arming is disarmed.
 
-registerMessageInterceptor(async (event: InboundEvent): Promise<boolean> => {
+const captureAgentNameReply = async (event: InboundEvent): Promise<boolean> => {
   const userId = extractAndUpsertUser(event);
   if (!userId) return false;
 
@@ -629,4 +641,20 @@ registerMessageInterceptor(async (event: InboundEvent): Promise<boolean> => {
     }
   }
   return true;
+};
+
+registerMessageInterceptor(captureAgentNameReply, {
+  action: channelsRegister,
+  claims: (event) => {
+    const userId = extractAndUpsertUser(event);
+    if (!userId) return null;
+    const pending = awaitingNameInput.get(userId);
+    if (!pending) return null;
+    if (event.channelType !== pending.dmChannelType || event.platformId !== pending.dmPlatformId) return null;
+    return { actor: { kind: 'human', userId }, payload: { questionId: pending.channelMgId } };
+  },
+  onDeny: (event) => {
+    const userId = extractAndUpsertUser(event);
+    if (userId) awaitingNameInput.delete(userId);
+  },
 });

src/modules/scheduling/index.ts

@@ -19,6 +19,7 @@
  * module piggybacks on the core schema.
  */
 import { registerDeliveryAction } from '../../delivery.js';
+import { unguarded } from '../../guard/index.js';
 import {
   handleCancelTask,
   handlePauseTask,
@@ -27,8 +28,10 @@ import {
   handleUpdateTask,
 } from './actions.js';
 
-registerDeliveryAction('schedule_task', handleScheduleTask);
-registerDeliveryAction('cancel_task', handleCancelTask);
-registerDeliveryAction('pause_task', handlePauseTask);
-registerDeliveryAction('resume_task', handleResumeTask);
-registerDeliveryAction('update_task', handleUpdateTask);
+const selfAction = unguarded('agent self-action — mutates only its own task rows; not a privileged class (yet)');
+
+registerDeliveryAction('schedule_task', handleScheduleTask, selfAction);
+registerDeliveryAction('cancel_task', handleCancelTask, selfAction);
+registerDeliveryAction('pause_task', handlePauseTask, selfAction);
+registerDeliveryAction('resume_task', handleResumeTask, selfAction);
+registerDeliveryAction('update_task', handleUpdateTask, selfAction);

src/modules/self-mod/apply.ts

@@ -1,11 +1,12 @@
 /**
- * Approval handlers for self-modification actions.
+ * Guarded handler bodies for self-modification actions.
  *
- * The approvals module calls these when an admin clicks Approve on a
- * pending_approvals row whose action matches. Each handler mutates the
- * container config in the DB, rebuilds/kills the container as needed,
- * and writes an on_wake message so the fresh container picks up where
- * the old one left off.
+ * The delivery registry's guard wrapper runs these only on `allow` — which,
+ * for self-mod, means an approved replay carrying a valid grant (the
+ * baseline holds unconditionally from the container path; see ./guard.ts).
+ * Each body mutates the container config in the DB, rebuilds/kills the
+ * container as needed, and writes an on_wake message so the fresh container
+ * picks up where the old one left off.
  *
  * install_packages: update DB + rebuild image + kill container + on_wake.
  * add_mcp_server: update DB + kill container + on_wake.
@@ -17,18 +18,19 @@ import { getSession } from '../../db/sessions.js';
 import type { McpServerConfig } from '../../container-config.js';
 import { log } from '../../log.js';
 import { writeSessionMessage } from '../../session-manager.js';
-import type { ApprovalHandler } from '../approvals/index.js';
+import type { Session } from '../../types.js';
+import { notifyAgent } from '../approvals/index.js';
 
-export const applyInstallPackages: ApprovalHandler = async ({ session, payload, userId, notify }) => {
+export async function applyInstallPackages(payload: Record<string, unknown>, session: Session): Promise<void> {
   const agentGroup = getAgentGroup(session.agent_group_id);
   if (!agentGroup) {
-    notify('install_packages approved but agent group missing.');
+    notifyAgent(session, 'install_packages approved but agent group missing.');
     return;
   }
 
   const configRow = getContainerConfig(agentGroup.id);
   if (!configRow) {
-    notify('install_packages approved but container config missing.');
+    notifyAgent(session, 'install_packages approved but container config missing.');
     return;
   }
 
@@ -52,7 +54,7 @@ export const applyInstallPackages: ApprovalHandler = async ({ session, payload,
     ...((payload.apt as string[] | undefined) || []),
     ...((payload.npm as string[] | undefined) || []),
   ].join(', ');
-  log.info('Package install approved', { agentGroupId: session.agent_group_id, userId });
+  log.info('Package install approved', { agentGroupId: session.agent_group_id });
   try {
     await buildAgentGroupImage(session.agent_group_id);
     writeSessionMessage(session.agent_group_id, session.id, {
@@ -75,23 +77,24 @@ export const applyInstallPackages: ApprovalHandler = async ({ session, payload,
     });
     log.info('Container rebuild completed (bundled with install)', { agentGroupId: session.agent_group_id });
   } catch (e) {
-    notify(
+    notifyAgent(
+      session,
       `Packages added to config (${pkgs}) but rebuild failed: ${e instanceof Error ? e.message : String(e)}. Tell the user — an admin will need to retry the install_packages request or inspect the build logs.`,
     );
     log.error('Bundled rebuild failed after install approval', { agentGroupId: session.agent_group_id, err: e });
   }
-};
+}
 
-export const applyAddMcpServer: ApprovalHandler = async ({ session, payload, userId, notify }) => {
+export async function applyAddMcpServer(payload: Record<string, unknown>, session: Session): Promise<void> {
   const agentGroup = getAgentGroup(session.agent_group_id);
   if (!agentGroup) {
-    notify('add_mcp_server approved but agent group missing.');
+    notifyAgent(session, 'add_mcp_server approved but agent group missing.');
     return;
   }
 
   const configRow = getContainerConfig(agentGroup.id);
   if (!configRow) {
-    notify('add_mcp_server approved but container config missing.');
+    notifyAgent(session, 'add_mcp_server approved but container config missing.');
     return;
   }
 
@@ -122,5 +125,5 @@ export const applyAddMcpServer: ApprovalHandler = async ({ session, payload, use
     const s = getSession(session.id);
     if (s) wakeContainer(s);
   });
-  log.info('MCP server add approved', { agentGroupId: session.agent_group_id, userId });
-};
+  log.info('MCP server add approved', { agentGroupId: session.agent_group_id });
+}

src/modules/self-mod/guard.ts

@@ -0,0 +1,32 @@
+/**
+ * Self-mod guard adapter — the module's catalog entries, composed at the
+ * module edge (imported by ./index.ts).
+ *
+ * The structural baseline is today's behavior verbatim: from the container
+ * path, self-modification is held unconditionally for the agent group's
+ * admin chain. (The equivalent host-side mutations — `ncl groups config
+ * add-package` etc. — are separate catalog actions derived from the command
+ * registry.)
+ */
+import { DENY, HOLD, defineGuardedAction, type GuardInput } from '../../guard/index.js';
+
+function selfModBaseline(label: string) {
+  return (input: GuardInput) => {
+    if (input.actor.kind !== 'agent') {
+      return DENY(`${label} is a container-originated action.`);
+    }
+    return HOLD(`${label} always requires admin approval from the container path`);
+  };
+}
+
+export const selfModInstallPackages = defineGuardedAction({
+  action: 'self_mod.install_packages',
+  approvalAction: 'install_packages',
+  baseline: selfModBaseline('install_packages'),
+});
+
+export const selfModAddMcpServer = defineGuardedAction({
+  action: 'self_mod.add_mcp_server',
+  approvalAction: 'add_mcp_server',
+  baseline: selfModBaseline('add_mcp_server'),
+});

src/modules/self-mod/index.ts

@@ -2,29 +2,51 @@
  * Self-modification module — admin-approved container mutations.
  *
  * Optional tier. Depends on the approvals default module for the request/
- * handler plumbing. On install the module registers:
- *   - Two delivery actions (install_packages, add_mcp_server) that validate
- *     input and queue an approval via requestApproval().
- *   - Two matching approval handlers that run on approve and perform the
- *     complete follow-up:
- *       install_packages → update container.json, rebuild image, kill
+ * handler plumbing and on the guard for the decision. On install the module
+ * registers:
+ *   - Its guard-catalog entries (./guard.ts): unconditional hold from the
+ *     container path.
+ *   - Two guard-wrapped delivery actions (install_packages, add_mcp_server):
+ *     validation runs as the wrapper's precheck, the hold builders card the
+ *     admin, and the handler bodies (./apply.ts) run only on allow — i.e. on
+ *     an approved replay:
+ *       install_packages → update container_configs, rebuild image, kill
  *         container (next wake respawns on the new image), schedule a
  *         verify-and-report follow-up prompt.
- *       add_mcp_server → update container.json, kill container. No image
+ *       add_mcp_server → update container_configs, kill container. No image
  *         rebuild — bun runs TS directly, so the new MCP server is wired
  *         by the next container start.
+ *   - Two approval handlers that re-enter the wrapped actions with the
+ *     approval row as the grant (one replay semantics — the guard re-checks
+ *     the structural baseline live).
  *
  * Without this module: the MCP tools in the container still write outbound
  * system messages with these actions, but delivery logs "Unknown system
  * action" and drops them. Admin never sees a card; nothing changes.
  */
-import { registerDeliveryAction } from '../../delivery.js';
-import { registerApprovalHandler } from '../approvals/index.js';
+import { reenterGuardedDeliveryAction, registerDeliveryAction } from '../../delivery.js';
+import { notifyAgent, registerApprovalHandler } from '../approvals/index.js';
 import { applyAddMcpServer, applyInstallPackages } from './apply.js';
-import { handleAddMcpServer, handleInstallPackages } from './request.js';
+import { selfModAddMcpServer, selfModInstallPackages } from './guard.js';
+import {
+  requestAddMcpServerHold,
+  requestInstallPackagesHold,
+  validateAddMcpServer,
+  validateInstallPackages,
+} from './request.js';
 
-registerDeliveryAction('install_packages', handleInstallPackages);
-registerDeliveryAction('add_mcp_server', handleAddMcpServer);
+registerDeliveryAction('install_packages', applyInstallPackages, {
+  guardAction: selfModInstallPackages,
+  precheck: validateInstallPackages,
+  requestHold: requestInstallPackagesHold,
+  onDeny: (_content, session, reason) => notifyAgent(session, `install_packages denied: ${reason}`),
+});
+registerDeliveryAction('add_mcp_server', applyAddMcpServer, {
+  guardAction: selfModAddMcpServer,
+  precheck: validateAddMcpServer,
+  requestHold: requestAddMcpServerHold,
+  onDeny: (_content, session, reason) => notifyAgent(session, `add_mcp_server denied: ${reason}`),
+});
 
-registerApprovalHandler('install_packages', applyInstallPackages);
-registerApprovalHandler('add_mcp_server', applyAddMcpServer);
+registerApprovalHandler('install_packages', reenterGuardedDeliveryAction('install_packages'));
+registerApprovalHandler('add_mcp_server', reenterGuardedDeliveryAction('add_mcp_server'));

src/modules/self-mod/request.ts

@@ -1,12 +1,12 @@
 /**
- * Delivery-action handlers for agent-initiated self-modification requests.
+ * Validation + hold-request builders for agent-initiated self-modification.
  *
  * Two actions the container can write into messages_out (via the self-mod
- * MCP tools): install_packages, add_mcp_server. Each one validates input
- * and queues an approval request. The admin's approval triggers the
- * matching approval handler in ./apply.ts, which also performs the
- * required follow-up (rebuild+restart for install_packages, restart-only
- * for add_mcp_server).
+ * MCP tools): install_packages, add_mcp_server. The delivery registry wraps
+ * each one with the guard (see ./guard.ts — unconditional hold from the
+ * container path): validation here runs as the wrapper's precheck, and the
+ * hold builders create the approval card when the guard holds. On approve,
+ * the continuation re-enters the wrapped action and ./apply.ts runs.
  *
  * Host-side sanitization for install_packages is defense-in-depth — the MCP
  * tool validates first. Both layers matter: the DB row carries the payload
@@ -17,40 +17,48 @@ import { log } from '../../log.js';
 import type { Session } from '../../types.js';
 import { notifyAgent, requestApproval } from '../approvals/index.js';
 
-export async function handleInstallPackages(content: Record<string, unknown>, session: Session): Promise<void> {
+export function validateInstallPackages(content: Record<string, unknown>, session: Session): boolean {
   const agentGroup = getAgentGroup(session.agent_group_id);
   if (!agentGroup) {
     notifyAgent(session, 'install_packages failed: agent group not found.');
-    return;
+    return false;
   }
 
   const apt = (content.apt as string[]) || [];
   const npm = (content.npm as string[]) || [];
-  const reason = (content.reason as string) || '';
 
   const APT_RE = /^[a-z0-9][a-z0-9._+-]*$/;
   const NPM_RE = /^(@[a-z0-9][a-z0-9._-]*\/)?[a-z0-9][a-z0-9._-]*$/;
   const MAX_PACKAGES = 20;
   if (apt.length + npm.length === 0) {
     notifyAgent(session, 'install_packages failed: at least one apt or npm package is required.');
-    return;
+    return false;
   }
   if (apt.length + npm.length > MAX_PACKAGES) {
     notifyAgent(session, `install_packages failed: max ${MAX_PACKAGES} packages per request.`);
-    return;
+    return false;
   }
   const invalidApt = apt.find((p) => !APT_RE.test(p));
   if (invalidApt) {
     notifyAgent(session, `install_packages failed: invalid apt package name "${invalidApt}".`);
     log.warn('install_packages: invalid apt package rejected', { pkg: invalidApt });
-    return;
+    return false;
   }
   const invalidNpm = npm.find((p) => !NPM_RE.test(p));
   if (invalidNpm) {
     notifyAgent(session, `install_packages failed: invalid npm package name "${invalidNpm}".`);
     log.warn('install_packages: invalid npm package rejected', { pkg: invalidNpm });
-    return;
+    return false;
   }
+  return true;
+}
+
+export async function requestInstallPackagesHold(content: Record<string, unknown>, session: Session): Promise<void> {
+  const agentGroup = getAgentGroup(session.agent_group_id);
+  if (!agentGroup) return;
+  const apt = (content.apt as string[]) || [];
+  const npm = (content.npm as string[]) || [];
+  const reason = (content.reason as string) || '';
 
   const packageList = [...apt.map((p) => `apt: ${p}`), ...npm.map((p) => `npm: ${p}`)].join(', ');
   await requestApproval({
@@ -63,18 +71,26 @@ export async function handleInstallPackages(content: Record<string, unknown>, se
   });
 }
 
-export async function handleAddMcpServer(content: Record<string, unknown>, session: Session): Promise<void> {
+export function validateAddMcpServer(content: Record<string, unknown>, session: Session): boolean {
   const agentGroup = getAgentGroup(session.agent_group_id);
   if (!agentGroup) {
     notifyAgent(session, 'add_mcp_server failed: agent group not found.');
-    return;
+    return false;
   }
   const serverName = content.name as string;
   const command = content.command as string;
   if (!serverName || !command) {
     notifyAgent(session, 'add_mcp_server failed: name and command are required.');
-    return;
+    return false;
   }
+  return true;
+}
+
+export async function requestAddMcpServerHold(content: Record<string, unknown>, session: Session): Promise<void> {
+  const agentGroup = getAgentGroup(session.agent_group_id);
+  if (!agentGroup) return;
+  const serverName = content.name as string;
+  const command = content.command as string;
   await requestApproval({
     session,
     agentName: agentGroup.name,

src/response-registry.ts

@@ -7,10 +7,21 @@
  * which triggers module registrations that would otherwise happen before
  * index.ts's own const initializers have run.
  *
- * Keep this file dependency-free (log.js is fine, but nothing from
- * modules/* or index.ts itself). Any file imported here must not in turn
- * import from src/index.ts, or the cycle returns.
+ * Keep this file dependency-free (log.js and the guard leaf are fine, but
+ * nothing from modules/* or index.ts itself). Any file imported here must
+ * not in turn import from src/index.ts, or the cycle returns.
+ *
+ * A handler whose click performs a privileged operation registers with a
+ * guard spec: the registry wraps it so the guard's decision stands between
+ * the click and the handler, and the wrapped path is the only path. `claims`
+ * is the handler's own claim test (does this questionId belong to me?) so an
+ * unauthorized click is claimed-and-dropped without stealing other handlers'
+ * responses. The guard argument is not optional — a handler that runs
+ * unguarded must declare so with `unguarded(<reason>)`, at the registration
+ * site, in the diff that adds it.
  */
+import { guard, type GuardActor, type GuardedAction, type Unguarded } from './guard/index.js';
+import { log } from './log.js';
 
 export interface ResponsePayload {
   questionId: string;
@@ -23,10 +34,46 @@ export interface ResponsePayload {
 
 export type ResponseHandler = (payload: ResponsePayload) => Promise<boolean>;
 
+export interface ResponseGuardSpec {
+  /** Guard action consulted before the handler runs — the defined value, not a name. */
+  action: GuardedAction;
+  /** Would this handler claim the response? (Its own row lookup.) */
+  claims: (payload: ResponsePayload) => boolean;
+}
+
 const responseHandlers: ResponseHandler[] = [];
 
-export function registerResponseHandler(handler: ResponseHandler): void {
-  responseHandlers.push(handler);
+function responseActor(payload: ResponsePayload): GuardActor {
+  if (!payload.userId) return { kind: 'human', userId: '' };
+  const userId = payload.userId.includes(':') ? payload.userId : `${payload.channelType}:${payload.userId}`;
+  return { kind: 'human', userId };
+}
+
+export function registerResponseHandler(handler: ResponseHandler, guardDecl: ResponseGuardSpec | Unguarded): void {
+  if ('reason' in guardDecl) {
+    // Explicitly declared unguarded — the carried reason is the reviewable record.
+    responseHandlers.push(handler);
+    return;
+  }
+  const spec = guardDecl;
+  responseHandlers.push(async (payload) => {
+    if (!spec.claims(payload)) return false;
+    const decision = guard(spec.action, {
+      actor: responseActor(payload),
+      payload: { questionId: payload.questionId, value: payload.value },
+    });
+    if (decision.effect !== 'allow') {
+      // Claim the response so it's not unclaimed-logged, but do nothing.
+      log.warn('Response click rejected by guard', {
+        action: spec.action.action,
+        questionId: payload.questionId,
+        userId: payload.userId,
+        reason: decision.reason,
+      });
+      return true;
+    }
+    return handler(payload);
+  });
 }
 
 export function getResponseHandlers(): readonly ResponseHandler[] {

src/router.ts

@@ -20,6 +20,7 @@
 import { getChannelAdapter } from './channels/channel-registry.js';
 import { gateCommand } from './command-gate.js';
 import { getAgentGroup } from './db/agent-groups.js';
+import { guard, type GuardActor, type GuardedAction, type Unguarded } from './guard/index.js';
 import { recordDroppedMessage } from './db/dropped-messages.js';
 import {
   createMessagingGroup,
@@ -117,13 +118,53 @@ export function setSenderScopeGate(fn: SenderScopeGateFn): void {
  * Used by modules to capture free-text DM replies during multi-step approval
  * flows — the permissions module (agent naming during channel registration)
  * and the approvals module (reject-with-reason capture).
+ *
+ * An interceptor whose capture performs a privileged operation (the
+ * channel-registration name capture creates an agent group + wiring)
+ * registers with a guard spec: the registry wraps it so the guard's decision
+ * stands between the free-text reply and the handler. `claims` returns the
+ * guard consult for events the interceptor would act on (null = not mine,
+ * pass through); a deny consumes the message without acting. The guard
+ * argument is not optional — an interceptor that runs unguarded must declare
+ * so with `unguarded(<reason>)` at the registration site.
  */
 export type MessageInterceptorFn = (event: InboundEvent) => Promise<boolean>;
 
+export interface InterceptorGuardSpec {
+  /** Guard action consulted before the interceptor acts — the defined value, not a name. */
+  action: GuardedAction;
+  /** The guard consult for events this interceptor would act on; null = not mine. */
+  claims: (event: InboundEvent) => { actor: GuardActor; payload: Record<string, unknown> } | null;
+  /** Domain cleanup when the guard denies (e.g. disarm the capture). */
+  onDeny?: (event: InboundEvent) => void;
+}
+
 const messageInterceptors: MessageInterceptorFn[] = [];
 
-export function registerMessageInterceptor(fn: MessageInterceptorFn): void {
-  messageInterceptors.push(fn);
+export function registerMessageInterceptor(
+  fn: MessageInterceptorFn,
+  guardDecl: InterceptorGuardSpec | Unguarded,
+): void {
+  if ('reason' in guardDecl) {
+    // Explicitly declared unguarded — the carried reason is the reviewable record.
+    messageInterceptors.push(fn);
+    return;
+  }
+  const guardSpec = guardDecl;
+  messageInterceptors.push(async (event) => {
+    const consult = guardSpec.claims(event);
+    if (!consult) return fn(event);
+    const decision = guard(guardSpec.action, { actor: consult.actor, payload: consult.payload });
+    if (decision.effect !== 'allow') {
+      log.warn('Interceptor capture rejected by guard — consuming without acting', {
+        action: guardSpec.action.action,
+        reason: decision.reason,
+      });
+      guardSpec.onDeny?.(event);
+      return true;
+    }
+    return fn(event);
+  });
 }
 
 /**