Install
$ agentstack add skill-ucsandman-dashclaw-skills-build-dashclaw ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Build DashClaw
Guide for contributing to the DashClaw codebase. Covers architecture, adding new capabilities, testing, and CI.
Architecture Overview
DashClaw is organized into 3 tiers:
Tier 1: Core Runtime (app/api/)
7 mandatory endpoints that define DashClaw's governance category:
| Route | Method | Purpose | |-------|--------|---------| | /api/guard | POST | Policy evaluation — "can I do this?" | | /api/actions | POST, PATCH, GET | Action lifecycle recording | | /api/approvals | POST | Human approval decisions | | /api/assumptions | POST, GET | Reasoning integrity tracking | | /api/signals | GET | Real-time anomaly detection | | /api/policies | GET, POST | Guard policy management | | /api/health | GET | System readiness |
The 7-route boundary is CI-enforced. Adding a new core route requires justification.
Tier 2: Extensions (app/(extensions)/)
Modular operational intelligence:
- Compliance, Drift, Evaluations, Scoring, Prompts, Webhooks, Learning
Tier 3: Archived (app/api/_archive/)
Legacy features from the "Agent Platform" era. Physically isolated.
Tech Stack
| Layer | Technology | |-------|------------| | Runtime | Node.js 20+ | | Framework | Next.js 16 (App Router) | | Database | Postgres (Neon recommended) | | ORM | Drizzle ORM | | Auth | NextAuth.js v4 | | UI | React 18, Tailwind CSS 3 | | Testing | Vitest + jsdom | | Package Manager | npm |
Key Conventions
TEXT Primary Keys
All IDs are TEXT with crypto-random prefixed values:
import { randomBytes } from 'crypto';
const id = `act_${randomBytes(16).toString('hex')}`;
ID prefixes: ar_ (actions), oc_live_/oc_test_ (API keys), sp_ (scoring profiles), pt_ (prompt templates), etc.
Repository Pattern
No direct SQL in route handlers. All data access goes through repository modules:
app/lib/repositories/*.repository.js
Route handlers import from repositories, never inline SQL.
Auth Chain
Request → Strip client org headers → Rate limit → Route matching
→ Public routes: pass through
→ Protected routes: x-api-key (fast path: timing-safe compare)
→ Inject org context from resolved key
12-Step Scaffold for New Capabilities
When adding a new feature to DashClaw:
- Migration — Add table with TEXT PKs and crypto-random IDs
- Repository — Create
app/lib/repositories/.repository.js(all SQL here) - Lib module — Create
app/lib/.js(business logic) - Route handler — Create
app/api//route.js(imports from repository) - Demo fixtures — Add fixture data if the feature needs demo mode support
- Demo middleware — Wire fixtures into demo mode if needed
- Node SDK method — Add to
sdk/dashclaw.js(camelCase) - Python SDK method — Add to
sdk-python/dashclaw/client.py(snake_case) - Docs page — Create dashboard page or update existing docs
- Node README — Add method to SDK README
- Python README — Add method to Python SDK README
- Parity matrix — Update
docs/sdk-parity.mdwith new method counts
Legacy SDK Promotion
The v1 SDK has 187 methods across 31 categories. When promoting to v2:
High priority (core governance):
registerOpenLoop()+resolveOpenLoop()— Decision integrityevents()— Real-time SSE eventsheartbeat()+startHeartbeat()— Agent telemetry
Medium priority (analytics):
getRecommendations()+recommendAction()— Adaptive learninggetLearningVelocity()+getLearningCurves()— Learning analyticsgetSignals()— Decision integrity signals
See knowledge/legacy-sdk-reference.md for the full inventory with all method signatures.
Testing
# Run all tests (watch mode)
npm run test
# Run tests once (CI mode)
npm run test -- --run
# Run specific test file
npx vitest run app/lib/__tests__/guard.test.js
Tests live alongside their modules or in __tests__/ directories.
CI Checks
Three mandatory CI checks:
# 1. Governance boundary — enforces 7-route API limit
npm run governance:boundary:check
# 2. OpenAPI contract — detects API drift
npm run openapi:check
# 3. Tests
npm run test -- --run
All three must pass before merge.
Key Files for Contributors
| File | What to Read | |------|-------------| | PROJECT_DETAILS.md | Canonical system map (source of truth) | | CLAUDE.md | Contributing conventions | | app/lib/guard.js | Risk scoring engine | | app/lib/signals.js | Anomaly detection logic | | sdk/dashclaw.js | V2 SDK implementation | | sdk/legacy/dashclaw-v1.js | V1 SDK (187 methods) | | schema/schema.js | Database schema (Drizzle) | | middleware.js | Auth chain | | docs/sdk-parity.md | SDK method parity matrix |
Contributing Workflow
- Read
PROJECT_DETAILS.mdandCLAUDE.md - Understand the 3-tier boundary (core vs extension vs archived)
- Follow the 12-step scaffold if adding a new capability
- Match existing patterns (repository pattern, TEXT PKs, etc.)
- Run all three CI checks locally before committing
- Keep SDK parity: if you add a Node method, add the Python equivalent
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: ucsandman
- Source: ucsandman/dashclaw-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.