Tools for agents that work. Built from real operation, not theory.

Keats AI builds practical systems for OpenClaw:

• reusable skills
• operational patterns
• agent foundations
• real workflows extracted from running agents in production

We care about one thing: making agents more useful, reliable, and governable in the real world.

• Persona Builder — guided interview that generates a production-grade agent identity and workspace foundation

• OpenForge — staged PRD execution for OpenClaw agents
• Repo Guardian — PR review governance and repository maintenance automation
• Adversarial Code Reviewer — rigorous, adversarial code review for code and PRs
• QA Gate — final quality validation before release or publication

• Frontend Design — distinctive, production-grade frontend design
• Design Engineering — design-heavy implementation orchestration and refinement
• Webapp Testing — browser-based testing and UI validation for local web apps

• Academic Deep Research — rigorous, transparent deep research with method and citations
• Web Monitor — watch specific pages for content changes and diffs

clawhub install persona-builder
clawhub install openforge

Browse the full public skill catalog here:

• keats-ai.dev/skills

We are building toward a more complete agent stack:

• identity and memory foundations
• operational hygiene and governance
• execution pipelines for real work
• research, monitoring, and validation loops
• productized operational playbooks

• Evidence before claims. Ship what survives contact with reality.
• Agents need operating structure. Identity, memory, trust, and governance matter.
• Reliability beats demos. The goal is sustained usefulness, not novelty theater.
• Open tools, sharp boundaries. Share what helps others build; keep private systems and product IP properly scoped.

• 🌐 Website: keats-ai.dev
• 📦 ClawHub publisher: corbin-breton on ClawHub
• 🦞 Platform: OpenClaw