23f7624596
ADR-166 MCP Bridge Security Lock / Static-source security lock (push) Failing after 0s
ADR-166 MCP Bridge Security Lock / Compose default binds loopback + Mongo has auth (push) Failing after 2s
CodeQL Advanced / Analyze (rust) (push) Failing after 0s
ADR-166 MCP Bridge Security Lock / plugin-agent-federation bindHost default (push) Failing after 1s
ADR-166 MCP Bridge Security Lock / Runtime behavior — 401 + terminal gate + fail-closed (push) Failing after 4s
business-pods-smoke / smoke (push) Failing after 1s
all-plugins-smoke / smoke-all (push) Failing after 2s
CI/CD Pipeline / Security & Code Quality (push) Failing after 1s
CI/CD Pipeline / Test Suite (ubuntu-latest) (push) Failing after 1s
CI/CD Pipeline / Build & Package (macos-latest) (push) Has been skipped
CI/CD Pipeline / Build & Package (ubuntu-latest) (push) Has been skipped
CI/CD Pipeline / Build & Package (windows-latest) (push) Has been skipped
CI/CD Pipeline / Documentation & Examples (push) Failing after 1s
Clone Tracker (14-day rolling) / Snapshot clones for ruflo ecosystem (push) Failing after 1s
CodeQL Advanced / Analyze (actions) (push) Failing after 1s
CodeQL Advanced / Analyze (javascript-typescript) (push) Failing after 1s
federation-peer-rust / stable-noop (push) Failing after 1s
metaharness-ci / score (push) Failing after 1s
metaharness-ci / router-compat (push) Failing after 0s
metaharness-ci / similarity-tests (push) Failing after 0s
no-agentbbs-smoke / smoke-without-agentbbs (push) Failing after 1s
V3 CI/CD Pipeline / Build V3 (windows-latest) (push) Has been skipped
codex-integration-audit / Codex integration audit (push) Failing after 1s
helpers-manifest-guard / guard (push) Failing after 1s
🔗 Cross-Agent Integration Tests / 🤝 Agent Coordination Tests (push) Has been skipped
🔗 Cross-Agent Integration Tests / 🧠 Memory Sharing Integration (push) Has been skipped
🔗 Cross-Agent Integration Tests / 🛡️ Fault Tolerance Tests (push) Has been skipped
🔗 Cross-Agent Integration Tests / ⚡ Performance Integration Tests (push) Has been skipped
metaharness-ci / mcp-scan (push) Failing after 1s
metaharness-ci / eject-dryrun (push) Failing after 1s
metaharness-ci / metaharness-real-data (push) Failing after 0s
no-cli-optdep-bloat-2561 / guard (push) Failing after 1s
no-metaharness-smoke / smoke-without-metaharness (push) Failing after 1s
no-phantom-agentic-flow-subpath / guard (push) Failing after 1s
🔄 Automated Rollback Manager / 🚨 Failure Detection (push) Failing after 1s
V3 CI/CD Pipeline / Plugin hooks smoke / ubuntu-latest / Node 22 (push) Failing after 1s
V3 CI/CD Pipeline / ruflo-graph-intelligence build + test smoke (#2044, ADR-123) (push) Failing after 1s
CVE Audit Gate / Audit root (critical-blocking) (push) Failing after 2s
cost-tracker-smoke / smoke (push) Failing after 3s
oia-audit-weekly / audit (push) Failing after 2s
ruflo-agent-smoke / ruflo-agent structural smoke (push) Failing after 1s
📊 Status Badges Update / 📊 Update Status Badges (push) Failing after 1s
V3 CI/CD Pipeline / Static regression guards (#2267 YAML + (push) Failing after 1s
V3 CI/CD Pipeline / Test V3 Packages (push) Failing after 0s
V3 CI/CD Pipeline / agent_execute provider routing smoke (#2042) (push) Failing after 0s
CVE Audit Gate / Audit v3 (critical-blocking) (push) Failing after 1s
federation-peer-rust / stable-native (push) Failing after 2s
🔗 Cross-Agent Integration Tests / 🚀 Integration Test Setup (push) Failing after 2s
neural-trader-smoke / runtime-smoke (push) Failing after 1s
V3 CI/CD Pipeline / Build V3 (macos-latest) (push) Has been skipped
V3 CI/CD Pipeline / Build V3 (ubuntu-latest) (push) Has been skipped
V3 CI/CD Pipeline / Type Check V3 (push) Failing after 1s
V3 CI/CD Pipeline / Smoke (no better-sqlite3) / ubuntu-latest / Node 24 (push) Failing after 1s
V3 CI/CD Pipeline / Smoke (no better-sqlite3) / ubuntu-latest / Node 22 (push) Failing after 2s
V3 CI/CD Pipeline / browser rvf create flag smoke (#2015) (push) Failing after 0s
V3 CI/CD Pipeline / Dependency review (#2046) (push) Has been skipped
V3 CI/CD Pipeline / Supply-chain audit (#2046) (push) Failing after 0s
V3 CI/CD Pipeline / witness marker drift smoke (#2021) (push) Failing after 1s
V3 CI/CD Pipeline / neural-trader portfolio CG smoke (#2068, ADR-126 Phase 3) (push) Failing after 1s
V3 CI/CD Pipeline / neural-trader backtest signing smoke (#2068, ADR-126 Phase 4) (push) Failing after 1s
V3 CI/CD Pipeline / kg-extract type-import classification smoke (#2049) (push) Failing after 0s
V3 CI/CD Pipeline / witness verify precondition smoke (#1880) (push) Failing after 2s
V3 CI/CD Pipeline / neural-trader pipeline risk-gate smoke (#2068, ADR-126 Phase 5) (push) Failing after 0s
V3 CI/CD Pipeline / neural-trader feature attribution smoke (#2068, ADR-126 Phase 6) (push) Failing after 0s
V3 CI/CD Pipeline / plugin-registry signature verification smoke (#1922, CWE-347) (push) Failing after 4s
V3 CI/CD Pipeline / memory stats legacy-DB smoke (#2120) (push) Failing after 4s
V3 CI/CD Pipeline / github deprecated actions smoke (#2089, ADR-127 Phase 3) (push) Failing after 1s
V3 CI/CD Pipeline / graph query + pathfinder smoke (ADR-130 P2+P5) (push) Has been skipped
V3 CI/CD Pipeline / graph trajectory hooks smoke (ADR-130 P3) (push) Has been skipped
V3 CI/CD Pipeline / graph plugin adapter smoke (ADR-130 P4) (push) Has been skipped
V3 CI/CD Pipeline / graph benchmark (ADR-130 P6) (push) Has been skipped
V3 CI/CD Pipeline / statusline generator delegation smoke (#2195) (push) Failing after 1s
V3 CI/CD Pipeline / wizard init regression guard (#2206 (push) Failing after 1s
V3 CI/CD Pipeline / memory no-stray-db smoke (ADR-125 P7) (push) Failing after 1s
V3 CI/CD Pipeline / github-safe injection smoke (#2089, ADR-127 Phase 1) (push) Failing after 1s
V3 CI/CD Pipeline / github actions pin smoke (#2089, ADR-127 Phase 1) (push) Failing after 1s
V3 CI/CD Pipeline / github attribution opt-in smoke (#2089, ADR-127 Phase 4) (push) Failing after 1s
V3 CI/CD Pipeline / pre-bash hook safety smoke (#2017) (push) Failing after 1s
V3 CI/CD Pipeline / Memory import smoke / ubuntu-latest (push) Failing after 0s
V3 CI/CD Pipeline / MCP protocol smoke / ubuntu-latest (push) Failing after 2s
V3 CI/CD Pipeline / ruvllm WASM auto-init smoke (#2086) (push) Failing after 4s
V3 CI/CD Pipeline / MCP paired-tool round-trip smoke (#1889) (push) Failing after 1s
V3 CI/CD Pipeline / Plugin package install-safety (#1902/#1903/#1904) (push) Failing after 1s
V3 CI/CD Pipeline / Tool description discoverability (ADR-112) (push) Failing after 3s
V3 CI/CD Pipeline / CLI npx-install smoke (#1147 / (22) (push) Failing after 1s
V3 CI/CD Pipeline / CLI npx-install smoke (#1147 / (24) (push) Failing after 1s
V3 CI/CD Pipeline / Windows hook shim smoke (#2132) / ubuntu-latest (push) Failing after 2s
V3 CI/CD Pipeline / Windows hook execution smoke (#2132) / ubuntu-latest (push) Failing after 1s
V3 CI/CD Pipeline / Windows init hooks smoke (#2132) / ubuntu-latest (push) Failing after 1s
V3 CI/CD Pipeline / Vector-index dimension audit (#1947) (push) Failing after 0s
V3 CI/CD Pipeline / Hook-command install safety (#1921) (push) Failing after 1s
V3 CI/CD Pipeline / ToolOutputGuardrail smoke (ADR-131, (push) Failing after 1s
V3 CI/CD Pipeline / init-bundle invariants smoke (#2095, ADR-128 Phase 5) (push) Failing after 1s
V3 CI/CD Pipeline / wasm provider bridge smoke (ADR-129 P1) (push) Failing after 2s
V3 CI/CD Pipeline / wasm gallery CRUD smoke (ADR-129 P3) (push) Failing after 1s
V3 CI/CD Pipeline / wasm plugin bridge smoke (ADR-129 P4) (push) Failing after 0s
V3 CI/CD Pipeline / wasm compose smoke (ADR-129 P2) (push) Failing after 4s
V3 CI/CD Pipeline / graph schema smoke (ADR-130 P1) (push) Failing after 0s
Validate Marketplace / validate (push) Failing after 1s
🔍 Verification Pipeline / 🚀 Setup Verification (push) Failing after 1s
🔍 Verification Pipeline / 🛡️ Security Verification (push) Has been skipped
🔍 Verification Pipeline / 📝 Code Quality (push) Has been skipped
🔍 Verification Pipeline / 🧪 Test Verification (${{ matrix.os }}, Node ${{ matrix.node }}) (push) Has been skipped
🔍 Verification Pipeline / 🏗️ Build Verification (push) Has been skipped
🔍 Verification Pipeline / 📚 Documentation Verification (push) Has been skipped
CVE Audit Gate / High-severity report (warn only) (push) Has been cancelled
🔄 Automated Rollback Manager / 🔄 Execute Rollback (push) Has been cancelled
🔄 Automated Rollback Manager / ✅ Post-Rollback Verification (push) Has been cancelled
🔄 Automated Rollback Manager / 📊 Rollback Monitoring (push) Has been cancelled
V3 CI/CD Pipeline / Windows init hooks smoke (#2132) / windows-latest (push) Has been cancelled
V3 CI/CD Pipeline / Windows hook execution smoke (#2132) / macos-latest (push) Has been cancelled
V3 CI/CD Pipeline / Windows hook execution smoke (#2132) / windows-latest (push) Has been cancelled
🔄 Automated Rollback Manager / ⏳ Manual Rollback Approval (push) Has been cancelled
V3 CI/CD Pipeline / MCP protocol smoke / macos-latest (push) Has been cancelled
V3 CI/CD Pipeline / Memory import smoke / macos-latest (push) Has been cancelled
V3 CI/CD Pipeline / Windows hook shim smoke (#2132) / macos-latest (push) Has been cancelled
V3 CI/CD Pipeline / Windows hook shim smoke (#2132) / windows-latest (push) Has been cancelled
V3 CI/CD Pipeline / Windows init hooks smoke (#2132) / macos-latest (push) Has been cancelled
V3 CI/CD Pipeline / Witness verify (signed manifest) / macos-latest (push) Has been cancelled
V3 CI/CD Pipeline / Witness verify (signed manifest) / ubuntu-latest (push) Has been cancelled
V3 CI/CD Pipeline / Witness verify (signed manifest) / windows-latest (push) Has been cancelled
V3 CI/CD Pipeline / Publish to npm (alpha) (push) Has been cancelled
V3 CI/CD Pipeline / Smoke (no better-sqlite3) / macos-latest / Node 22 (push) Has been cancelled
V3 CI/CD Pipeline / Plugin hooks smoke / macos-latest / Node 22 (push) Has been cancelled
CI/CD Pipeline / Deploy & Release (push) Has been cancelled
CI/CD Pipeline / CI Status (push) Has been cancelled
🔗 Cross-Agent Integration Tests / 📊 Integration Test Report (push) Has been cancelled
🔄 Automated Rollback Manager / 🔍 Pre-Rollback Validation (push) Has been cancelled
🔍 Verification Pipeline / ⚡ Performance Verification (push) Has been cancelled
🔍 Verification Pipeline / 📊 Verification Report (push) Has been cancelled
236 lines
8.4 KiB
Markdown
236 lines
8.4 KiB
Markdown
# Getting Started with @claude-flow/guidance
|
|
|
|
## Background: How Claude Code Uses Memory Files
|
|
|
|
In Claude Code, `CLAUDE.md` and `CLAUDE.local.md` are loaded into the agent's working context at session start. They are the primary mechanism for telling Claude Code how to behave on your project:
|
|
|
|
| File | Who it's for | Committed? |
|
|
|------|-------------|-----------|
|
|
| `CLAUDE.md` | The whole team | Yes |
|
|
| `CLAUDE.local.md` | You, on this machine | No (auto-gitignored) |
|
|
|
|
Claude Code searches **upward** from the current directory and loads every instance it finds. In a monorepo, a child directory's `CLAUDE.md` layers on top of the root's. Run `/memory` in Claude Code to see which files were loaded.
|
|
|
|
The `@import` pattern offers an alternative to `CLAUDE.local.md` for developers using multiple git worktrees:
|
|
|
|
```markdown
|
|
# In CLAUDE.md (committed):
|
|
@~/.claude/my_project_instructions.md
|
|
```
|
|
|
|
## What This Package Adds
|
|
|
|
The Guidance Control Plane takes these plain-text files and turns them into structured, enforceable, auditable policy. Without it, Claude Code loads `CLAUDE.md` as raw text and relies on the model to follow it. With it, rules become typed objects with gates that block violations before they execute.
|
|
|
|
## Installation
|
|
|
|
```bash
|
|
npm install @claude-flow/guidance@v3alpha
|
|
```
|
|
|
|
Requires Node.js 20+.
|
|
|
|
## Minimal Setup
|
|
|
|
```ts
|
|
import { createGuidanceControlPlane } from '@claude-flow/guidance';
|
|
|
|
const plane = createGuidanceControlPlane({
|
|
rootGuidancePath: './CLAUDE.md',
|
|
localGuidancePath: './CLAUDE.local.md', // optional, your personal overrides
|
|
});
|
|
await plane.initialize();
|
|
```
|
|
|
|
This reads both files, compiles them into a policy bundle (constitution + rule shards + manifest), and prepares all subsystems.
|
|
|
|
## What Happens at Initialization
|
|
|
|
1. **Load** — `CLAUDE.md` is read as the root guidance. `CLAUDE.local.md` (if present) is read as an overlay. Local rules supplement or override root rules.
|
|
2. **Compile** — Both files are parsed into structured rules. Each rule gets an ID, risk class, domain tags, tool class tags, and intent tags.
|
|
3. **Shard** — Rules are broken into task-scoped shards. The always-loaded invariants form the constitution (first ~30-60 lines of the root file). Everything else becomes retrievable shards.
|
|
4. **Index** — Shards are loaded into the retriever for similarity-based lookup.
|
|
5. **Activate gates** — Enforcement gates are configured from the compiled rules.
|
|
|
|
## Core Loop
|
|
|
|
Every agent task follows this pattern:
|
|
|
|
```ts
|
|
// 1. Retrieve relevant rules for this task
|
|
const guidance = await plane.retrieveForTask({
|
|
taskDescription: 'Fix the login timeout bug',
|
|
intent: 'bug-fix',
|
|
});
|
|
// guidance.constitution — always-loaded invariants
|
|
// guidance.shards — task-relevant rules
|
|
|
|
// 2. Gate commands before execution
|
|
const gateResults = plane.evaluateCommand('git reset --hard');
|
|
for (const result of gateResults) {
|
|
if (result.decision === 'deny') {
|
|
console.error(`Blocked: ${result.reason}`);
|
|
// Don't execute the command
|
|
}
|
|
}
|
|
|
|
// 3. Track the run
|
|
const event = plane.startRun('task-123', 'bug-fix');
|
|
// ... agent does work ...
|
|
const evaluations = await plane.finalizeRun(event);
|
|
```
|
|
|
|
## Using Individual Modules
|
|
|
|
You don't have to use the all-in-one control plane. Each module is independently importable:
|
|
|
|
```ts
|
|
// Just the gates
|
|
import { createGates } from '@claude-flow/guidance/gates';
|
|
const gates = createGates({ destructiveOps: true, secrets: true });
|
|
|
|
// Just the proof chain
|
|
import { createProofChain } from '@claude-flow/guidance/proof';
|
|
const chain = createProofChain('my-hmac-key');
|
|
|
|
// Just the trust system
|
|
import { createTrustSystem } from '@claude-flow/guidance/trust';
|
|
const trust = createTrustSystem();
|
|
```
|
|
|
|
All 20 modules are available as separate entry points. See the [API Reference](../reference/api-quick-reference.md) for the full list.
|
|
|
|
## CLAUDE.md vs. CLAUDE.local.md
|
|
|
|
### What goes in CLAUDE.md (shared, committed)
|
|
|
|
Team-level guidance everyone benefits from:
|
|
|
|
```markdown
|
|
# Build & Test
|
|
Run `npm test` before committing. Run `npm run build` to type-check.
|
|
|
|
# Coding Standards
|
|
- No `any` types. Use `unknown` when the type is truly unknown.
|
|
- All public functions require JSDoc.
|
|
|
|
# Architecture
|
|
This project uses a layered architecture. See docs/architecture.md.
|
|
|
|
# Domain Rules
|
|
- Never write to the `users` table without a migration.
|
|
- API responses must include `requestId` for tracing.
|
|
```
|
|
|
|
### What goes in CLAUDE.local.md (private, not committed)
|
|
|
|
Machine-specific or personal notes:
|
|
|
|
```markdown
|
|
# My Environment
|
|
- Local API: http://localhost:3001
|
|
- Test DB: postgres://localhost:5432/myapp_test
|
|
|
|
# Preferences
|
|
- Show git diffs before committing
|
|
- I prefer verbose error messages
|
|
```
|
|
|
|
### The @import alternative for worktrees
|
|
|
|
If you use multiple git worktrees, `CLAUDE.local.md` is awkward because each worktree needs its own copy. The `@` import pattern inside `CLAUDE.md` points to a shared file in your home directory:
|
|
|
|
```markdown
|
|
@~/.claude/my_project_instructions.md
|
|
```
|
|
|
|
### How the optimizer uses both files
|
|
|
|
The optimizer watches which `CLAUDE.local.md` experiments reduce violations. When a local rule consistently outperforms the root, the optimizer proposes promoting it to `CLAUDE.md` with an ADR:
|
|
|
|
```ts
|
|
const optimized = await plane.optimize();
|
|
// optimized.promoted — rules moved from local to root
|
|
// optimized.demoted — ineffective root rules flagged
|
|
// optimized.adrsCreated — decision records
|
|
```
|
|
|
|
## WASM Acceleration
|
|
|
|
If the pre-built WASM binary is available, hot-path operations (hashing, secret scanning, destructive detection) run 1.25-1.96x faster automatically:
|
|
|
|
```ts
|
|
import { getKernel, isWasmAvailable } from '@claude-flow/guidance/wasm-kernel';
|
|
|
|
console.log(isWasmAvailable()); // true if WASM loaded
|
|
const k = getKernel(); // WASM or JS fallback — same API either way
|
|
```
|
|
|
|
No configuration needed. The bridge detects WASM availability at load time.
|
|
|
|
## File Organization
|
|
|
|
```
|
|
@claude-flow/guidance/
|
|
src/
|
|
index.ts # Control plane + re-exports
|
|
compiler.ts # CLAUDE.md → PolicyBundle
|
|
retriever.ts # Shard similarity retrieval
|
|
gates.ts # Enforcement gates (4 built-in)
|
|
gateway.ts # Tool gateway (idempotency + schema + budget)
|
|
proof.ts # Hash-chained proof envelopes
|
|
continue-gate.ts # Step-level agent control
|
|
memory-gate.ts # Memory write authorization
|
|
capabilities.ts # Typed permission algebra
|
|
trust.ts # Trust score accumulation
|
|
authority.ts # Authority levels + irreversibility
|
|
adversarial.ts # Threat/collusion detection + quorum
|
|
meta-governance.ts # Governance over governance
|
|
coherence.ts # Coherence scoring + economic budgets
|
|
uncertainty.ts # Probabilistic belief tracking
|
|
temporal.ts # Bitemporal assertions
|
|
truth-anchors.ts # Immutable external facts
|
|
ledger.ts # Run logging + evaluators
|
|
optimizer.ts # Rule evolution
|
|
headless.ts # Automated compliance testing
|
|
wasm-kernel.ts # WASM host bridge
|
|
wasm-kernel/ # Rust source for WASM kernel
|
|
wasm-pkg/ # Pre-built WASM binary
|
|
tests/ # 1088 tests across 24 files
|
|
docs/
|
|
guides/ # Conceptual guides
|
|
tutorials/ # Step-by-step walkthroughs
|
|
reference/ # API reference
|
|
diagrams/ # Architecture diagrams
|
|
adrs/ # Architecture Decision Records (G001-G025)
|
|
```
|
|
|
|
## Verification
|
|
|
|
To confirm both files are being loaded and enforced:
|
|
|
|
1. Add a unique rule to `CLAUDE.md`: `# Test: Always respond in English`
|
|
2. Add a different rule to `CLAUDE.local.md`: `# Test: Prefer bullet points`
|
|
3. Start Claude Code and run `/memory` — both files should appear as loaded
|
|
4. Ask Claude to restate each unique rule to confirm both were applied
|
|
|
|
Then, initialize the guidance control plane and confirm:
|
|
|
|
```ts
|
|
const plane = createGuidanceControlPlane({
|
|
rootGuidancePath: './CLAUDE.md',
|
|
localGuidancePath: './CLAUDE.local.md',
|
|
});
|
|
await plane.initialize();
|
|
const bundle = plane.getBundle();
|
|
console.log(bundle.constitution.length); // > 0
|
|
console.log(bundle.shards.length); // > 0
|
|
```
|
|
|
|
## Next Steps
|
|
|
|
- [Architecture Overview](./architecture-overview.md) — How the 7 layers connect
|
|
- [Enforcement Gates Tutorial](../tutorials/enforcement-gates.md) — Wire gates into hooks
|
|
- [Multi-Agent Security](./multi-agent-security.md) — Threat detection, collusion, quorum
|
|
- [WASM Kernel Guide](./wasm-kernel.md) — Building and benchmarking
|