Open Source · ClawHub
Skills
Same research foundation as Blueprint. Install any skill in one command — useful immediately, composable with your existing agent from day one.
Browse 10 production skills. Each one is designed to stand alone, install in seconds, and compose with your existing setup.
OpenForge
PRD-to-implementation pipeline for OpenClaw. Multi-model execution with staged phases, escalation logic, and quality gates.
- Structured PRD execution workflow
- Parallel phases where possible
- Model routing and escalation
- Review-to-fix loops and validation gates
Repo Guardian
GitHub PR review governance and repository maintenance automation. Enforces merge hygiene and review discipline for active repos.
- PR governance and merge discipline
- Review gate automation
- Repository maintenance workflows
- Dual-model review-oriented operations
Persona Builder
Guided interview that generates a complete production-grade agent workspace with identity, memory, and operating files.
- Guided persona interview flow
- Generates core workspace identity files
- Research-backed memory structure
- Fast path from blank agent to working system
Adversarial Code Reviewer
Rigorous, zero-tolerance code review with severity tiers. Supports optional dual-model mode for cross-model iterative review on high-stakes PRs.
- Severity-tiered feedback: Blocking → Suggested
- Security holes, edge cases, lazy patterns
- Optional dual-model adversarial mode
- Python, JS/TS, SQL, and frontend
Frontend Design
Create distinctive, production-grade frontend interfaces. Anti-AI-slop aesthetics — avoids generic, templated output.
- Distinctive, non-generic UI output
- React, HTML/CSS, and component design
- Dark-first with light mode support
- Accessibility-conscious patterns
QA Gate
Final quality validation across 6 dimensions before human review. Factual accuracy, tone, completeness, structure, operational soundness, and sensitive data.
- 6-dimension validation framework
- Factual accuracy and tone consistency
- Sensitive data detection
- Works on docs, skills, PRDs, code
Design Engineering
Orchestrates iterative design and frontend engineering through research, planning, sub-agent execution, and validation loops.
- Research → plan → execute → validate loop
- Canvas vs SVG vs CSS decision-making
- Performance-aware rendering choices
- Sub-agent coordination for complex work
Webapp Testing
Playwright-based local webapp testing toolkit. Verifies frontend functionality, captures screenshots, debugs UI behavior — no manual browser required.
- Playwright automation (no manual browser)
- Screenshots and browser log capture
- Frontend functionality verification
- UI behavior debugging
Web Monitor
Monitors web pages for content changes and sends alerts. CSS selector support for targeted monitoring of specific page sections.
- CSS selector-targeted monitoring
- Diff view on content changes
- Alert delivery on detection
- Multi-URL tracking support
Deep Research
Academic-grade research with APA 7th citations and mandatory dual research cycles per theme. Full methodology transparency with evidence hierarchy.
- APA 7th edition citations
- Mandatory dual-cycle per theme
- Evidence hierarchy enforcement
- 3 user checkpoints for transparency
No skills in this category yet.
Want the complete system?
Skills run deeper with Blueprint.
These skills share the same memory and safety architecture as Blueprint. The full foundation — persistent memory, scheduling, trust ladder, self-improvement — is in Blueprint. That's where the architecture becomes a system.