Files
gbrain/recipes/agent-voice/code/tools.mjs
T
e9fa51d46e v0.40.0.0 feat: agent-voice (Mars + Venus) + copy-into-host-repo skillpack paradigm (#1128)
* feat: agent-voice reference skillpack (Mars + Venus) + copy-into-host-repo install paradigm

Ships a new skillpack paradigm: gbrain holds the REFERENCE content;
`gbrain integrations install agent-voice --target <repo>` COPIES it into
the operator's host agent repo where it becomes user-owned and mutable.
Future refresh is diff-and-propose against per-file SHA-256 hashes from
.gbrain-source.json, not blind overwrite.

What ships:
- recipes/agent-voice.md entrypoint + recipes/agent-voice/ bundle
- Two voice personas (Mars dual-mode SOLO/DEMO, Venus executive assistant)
  with PII / private-agent-name / hardcoded-path scrubbed out
- WebRTC-first browser client (call.html) with ?test=1 gated instrumentation
  and Web Audio API tee -> MediaRecorder capture for E2E roundtrip testing
- Read-only tool router (D14-A allow-list: search, query, get_page,
  list_pages, find_experts, get_recent_salience, get_recent_transcripts,
  read_article). Write ops permanently denylisted; opt-in via local override
- Persona-aware prompt builder with identity-first composition + Unicode
  sanitization for OpenAI Realtime API safety
- Upstream-error classifier (HTTP 429/500/503 -> soft-fail, plumbing -> hard)
- Three SKILL.md skills (voice-persona-mars, voice-persona-venus,
  voice-post-call) with routing-eval.jsonl fixtures
- 99 host-side tests (vitest-compatible, runs in bun) covering registry,
  prompt-shape privacy guards, tool allow-list, upstream classifier
- install/manifest.json + refresh-algorithm.md + post-install-hint.md

Privacy infrastructure:
- scripts/check-no-pii-in-agent-voice.sh wired into bun run verify
  Shape regex (phone/email/SSN/JWT/bearer/credit-card) + path patterns +
  $AGENT_VOICE_PII_BLOCKLIST env-driven name blocklist
- scripts/import-from-upstream.sh + scripts/upstream-scrub-table.txt
  Deterministic refresh from upstream voice-agent source. Placeholder-
  driven (envsubst-expanded at run time) so no private names land in
  checked-in files
- recipes/agent-voice/code/lib/personas/private-name-blocklist.json
  Single source of truth for the regex contract (shape categories +
  path patterns + env-var contract for operator-specific names)

src/ surface:
- src/commands/integrations.ts gains `install <recipe-id>` subcommand
  with install_kind: 'local-managed' | 'copy-into-host-repo' discriminator.
  Path-traversal hardening (rejects '..', absolute paths, symlink escapes).
  Refuses target == gbrain itself, missing .git, existing files (without
  --overwrite). Writes .gbrain-source.json with per-file SHA-256. Appends
  resolver rows to host repo's RESOLVER.md or AGENTS.md.
- test/integrations-install.test.ts: 11 cases (happy path, manifest shape,
  no upstream_repo field per D11-A, resolver appending, file modes,
  refusal cases, dry-run)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* chore: bump version and changelog (v0.36.0.0)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix(privacy): scrub literal private agent names from prompt-shape tests + guard script

The prompt-shape tests carried regex patterns naming the literal banned terms
(Garry/Steph/Garrison/Solomon/Herbert/Wintermute) inline. CLAUDE.md's
"never use Wintermute in any public artifact" applies to test source files
too. Master's check-privacy.sh correctly caught this.

Replaced with env-driven check that reads AGENT_VOICE_PII_BLOCKLIST (the
single source of truth from private-name-blocklist.json). Same enforcement
guarantee via the env var, zero literal names in shipped source.

Also scrubbed the literal /data/.openclaw/ from the guard script's comment
and the literal 'tell_wintermute' from the venus write-tools test.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* feat: ship all v0.36.0.0 deferred items in this PR (E2E + evals + pipeline + refresh + multilingual + twilio deprecation)

Closes the "deferred to follow-up" section of the v0.36.0.0 CHANGELOG.

E2E tests + harness (env-gated):
- tests/e2e/voice-roundtrip.test.mjs — spawns server, drives puppeteer + fake-audio, three-tier assertions (CONNECTION hard, NON-SILENT hard, SEMANTIC soft via Whisper + LLM judge). Upstream errors (429/500/503, WS 1011/1013) soft-fail via lib/upstream-classifier.mjs.
- tests/e2e/voice-full-flow.test.mjs — wraps openclaw doing the install, then runs the roundtrip. Friction-discovery flavor, NOT a ship gate.
- tests/e2e/lib/browser-audio.mjs — puppeteer + fake-audio harness; reads window._gbrainTest namespace; PCM RMS-variance helper.
- tests/e2e/lib/whisper-judge.mjs — Whisper transcription + LLM-judge for SEMANTIC tier.
- tests/e2e/audio-fixtures/utterance-{add,joke,brain-query}.wav — 16kHz mono WAV via `say` + ffmpeg, committed for reproducibility.
- test/fixtures/claw-test-scenarios/voice-agent-install/{BRIEF.md, scenario.json, expected.json} — labeled BENCHMARK_FRICTION, blocks_ship=false.

LLM-judge persona evals + synthetic canonical baselines:
- tests/evals/judge.mjs — gateway-routed 3-model (Claude + GPT + Gemini) harness with 4-strategy JSON repair + 2/3-quorum aggregation (per the v0.27.x cross-modal pattern). Pass criterion: every axis mean ≥7 AND no model <5.
- tests/evals/fixtures/{mars-solo,mars-demo,venus,persona-routing,mars-multilingual}.jsonl — 5 fixture sets covering all axes.
- tests/evals/{mars-eval,venus-eval,mars-multilingual-eval,persona-routing-eval}.mjs — per-axis drivers.
- tests/evals/baseline-runs/canonical/*.json — agent-authored synthetic exemplars (PII-impossible by construction; demonstrate expected pass shape; never overwrite with live model output).
- tests/evals/baseline-runs/.gitignore — live receipts excluded.

DIY pipeline (Option B):
- code/pipeline.mjs — streaming STT (Deepgram nova-2) + LLM (Claude Sonnet 4.6 streaming SSE with sentence-boundary TTS dispatch) + TTS (Cartesia primary, OpenAI TTS fallback). 20-turn history cap, exponential-backoff reconnects, 25s keepalives, VAD presets (quiet/normal/noisy/very_noisy), barge-in via STT speechStart → LLM interrupt. Modular adapters for swapping providers.

--refresh mode (D3-A diff-and-propose):
- src/commands/integrations.ts: refreshRecipeIntoHostRepo() + classifyForRefresh() implementing the five states from refresh-algorithm.md (unchanged-identical, unchanged-stale, locally-modified, source-deleted, host-deleted, new-in-manifest). Transaction journal at .gbrain-source.refresh.log. Default policy: preserve operator's local edits (keep-mine); --auto take-theirs to overwrite; --dry-run for preview.
- test/integrations-install.test.ts: 7 new test cases pinning each classification state + default-preserve behavior + take-theirs overwrite + transaction journal + refusal on uninstalled target.

Mars multilingual restore:
- code/lib/personas/mars.mjs: explicit cross-lingual rule (Mandarin, Spanish, French, Japanese, Korean default to English but follow the speaker). Voice (Orus) supports the languages natively.
- tests/unit/mars-prompt-shape.test.mjs: assertion flipped from "MUST NOT claim multilingual" to "declares cross-lingual capability with English bias."
- tests/evals/fixtures/mars-multilingual.jsonl: 5 fixtures across Mandarin/Spanish/Japanese/French + explicit switch-back, pinned by mars-multilingual-eval.mjs.

Twilio recipe deprecation:
- recipes/twilio-voice-brain.md: deprecation banner pointing at agent-voice.md. Frontmatter version bumped to 0.8.2. Will be removed in v0.37.

Verify: bun run verify clean, 6736+ unit tests pass, 18/18 install+refresh tests pass, 96/98 host-side persona/tool/classifier tests pass (2 skipped env-gated).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* docs: update CHANGELOG — all v0.36.0.0 deferred items now shipped in this PR

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* chore: rebump version v0.36.0.0 → v0.37.0.0

Captures the wave-1 + wave-2 scope at the v0.37 slot. The bump reflects
the size of what this PR ships: copy-into-host-repo install paradigm
(new install_kind discriminator + new install/refresh subcommand) +
Mars/Venus voice agent reference + 5,500+ LOC of vendored scrubbed
code + 4 LLM-judge eval suites + 2 env-gated E2E test suites + DIY
Option B pipeline + 18-case install subcommand test coverage. A minor
bump felt too small.

Side fix: privacy guard caught two stale literal "wintermute" and
"/data/.openclaw/" references in wave-2 files
(voice-full-flow.test.mjs comment, expected.json blocklist payload).
Both replaced with env-driven references to $AGENT_VOICE_PII_BLOCKLIST
matching the D15-A pattern from the original review.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* chore: rebump v0.37.0.0 → v0.40.0.0

Jumps past the v0.37/v0.38/v0.39 slots master might claim in subsequent
PRs. The wave's scope (copy-into-host-repo skillpack paradigm + agent-voice
+ install/refresh + LLM-judge evals + DIY pipeline + Mars multilingual)
justifies a larger version arithmetic step.

Files bumped:
- VERSION 0.37.0.0 → 0.40.0.0
- package.json 0.37.0.0 → 0.40.0.0
- CHANGELOG.md header + "To take advantage of v0.40.0.0" block
- recipes/twilio-voice-brain.md deprecation banner (now "removed in v0.41")
- recipes/agent-voice/tests/evals/mars-multilingual-eval.mjs comment

Left alone: master's pre-existing "v0.37+" roadmap labels in src/core/calibration/*,
src/core/cycle/*, DESIGN.md, CLAUDE.md, etc. Those are master's author-intent
references to "the next planned release" relative to master's frame at the time —
rewriting them just to keep numbering consistent would overreach.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-22 21:54:13 -07:00

176 lines
5.9 KiB
JavaScript

/**
* tools.mjs — voice-callable tool router with a READ-ONLY allow-list.
*
* D14-A invariant: the voice agent can READ from the brain but cannot WRITE.
* Prompt injection of a persona prompt cannot escalate to brain-mutating ops
* (put_page, submit_job, file_upload, delete_page, etc.) without explicit
* operator opt-in via a local `tools-allowlist.local.json` override.
*
* Allow-list (default):
* - search (semantic + keyword search over pages/chunks)
* - query (RAG-shaped query with citations)
* - get_page (fetch a single page by slug)
* - list_pages (list pages with filters)
* - find_experts (people/companies with topical relevance)
* - get_recent_salience (recently-active emotionally-loaded pages)
* - get_recent_transcripts (voice-note / meeting transcripts)
* - read_article (fetch + summarize an arbitrary URL)
*
* Reject path: any tool name outside the allow-list returns a structured
* error envelope. The voice agent surfaces it as "I can't do that from
* voice." The router NEVER raises an exception that crashes the call.
*
* Local override: an operator who wants voice-callable writes places a
* `tools-allowlist.local.json` next to this file with shape:
* { "extend": ["set_reminder", "log_to_brain"] }
* Only operations declared in OPTIONAL_OPS below can be added — wholesale
* write access (put_page, submit_job, etc.) is intentionally NOT
* extensible from the override.
*/
import { existsSync, readFileSync } from 'node:fs';
import { fileURLToPath } from 'node:url';
import { dirname, join } from 'node:path';
import { callGbrainOp } from './gbrain-client.mjs';
const __dirname = dirname(fileURLToPath(import.meta.url));
// Default read-only allow-list. KEEP IN SYNC with tests/unit/tools-allowlist.test.mjs.
export const READ_ONLY_OPS = Object.freeze([
'search',
'query',
'get_page',
'list_pages',
'find_experts',
'get_recent_salience',
'get_recent_transcripts',
'read_article',
]);
// Optional ops the operator MAY opt in to via tools-allowlist.local.json.
// These are write-shaped but bounded: a reminder, a brain-log, an enrich
// nudge. Crucially: not put_page (no arbitrary brain edits), not submit_job
// (no shell), not file_upload, not delete_page.
export const OPTIONAL_OPS = Object.freeze([
'set_reminder',
'log_to_brain',
'enrich_request',
]);
// Permanent denylist — never extendable via override, regardless of what the
// override file says. These cross the trust boundary in ways that would
// allow a prompt-injection attack to escalate to RCE or arbitrary write.
export const DENYLIST = Object.freeze([
'put_page',
'submit_job',
'file_upload',
'delete_page',
'file_url',
'sync_brain',
'apply_migrations',
'work',
'shell',
]);
/**
* Read the local override file (if present) and compute the effective
* allow-list. The override may add OPTIONAL_OPS but cannot add DENYLIST ops.
*
* Result is memoized per-process (the override file is read once at first
* call). Reset for tests by passing `{forceReload: true}`.
*/
let _effective;
export function getEffectiveAllowlist({ forceReload = false } = {}) {
if (_effective && !forceReload) return _effective;
const allowed = new Set(READ_ONLY_OPS);
const overridePath = join(__dirname, 'tools-allowlist.local.json');
if (existsSync(overridePath)) {
try {
const override = JSON.parse(readFileSync(overridePath, 'utf8'));
if (Array.isArray(override?.extend)) {
for (const name of override.extend) {
if (typeof name !== 'string') continue;
if (DENYLIST.includes(name)) {
console.warn(`[tools] ignoring denylisted op in override: ${name}`);
continue;
}
if (!OPTIONAL_OPS.includes(name)) {
console.warn(`[tools] ignoring unknown op in override: ${name}`);
continue;
}
allowed.add(name);
}
}
} catch (err) {
console.warn(`[tools] failed to parse ${overridePath}: ${err.message}`);
}
}
_effective = Object.freeze([...allowed]);
return _effective;
}
/**
* Structured error envelope returned to the voice persona when a tool is
* rejected. The persona sees this as a tool-call result, not as a thrown
* exception, so the call never hangs.
*/
export function rejectedToolResult(opName, reason) {
return {
error: {
code: reason,
op: opName,
message:
reason === 'denylisted'
? `Tool "${opName}" is permanently disabled in voice agents. Use a different surface.`
: `Tool "${opName}" is not in the voice agent's allow-list. The operator can opt in via tools-allowlist.local.json if appropriate.`,
},
};
}
/**
* Dispatch a tool call from the voice persona to gbrain MCP.
*
* Returns either {data: <result>} on success or {error: <envelope>} on
* rejection or MCP failure. Never throws.
*/
export async function dispatchTool(opName, params, { forceReload = false } = {}) {
// Denylist check first — even if someone added it to the allow-list,
// denylisted ops are never callable from voice.
if (DENYLIST.includes(opName)) {
return rejectedToolResult(opName, 'denylisted');
}
const allowlist = getEffectiveAllowlist({ forceReload });
if (!allowlist.includes(opName)) {
return rejectedToolResult(opName, 'not_in_allowlist');
}
// Allow-listed — call gbrain MCP.
try {
const result = await callGbrainOp(opName, params);
return { data: result };
} catch (err) {
return {
error: {
code: 'mcp_error',
op: opName,
message: err?.message || String(err),
},
};
}
}
/**
* Inspect helper for tests / introspection. Returns the full picture.
*/
export function describeAllowlist({ forceReload = false } = {}) {
return {
read_only: [...READ_ONLY_OPS],
optional: [...OPTIONAL_OPS],
denylist: [...DENYLIST],
effective: getEffectiveAllowlist({ forceReload }),
};
}