AI development toolkit with 11 specialized agents and 20 commands per tool. Supports Claude, Opencode, Ampcode, and Droid.
See README.md for usage and docs/INSTALLER_GUIDE.md for installation details.
Simple > clever. Readable code a junior can follow beats elegant code that needs a PhD to debug. Be the Simplicity Advocate: call out overcomplications and suggest simpler alternatives BEFORE building, not after.
Every line must have a purpose. No speculative code, no "might need this later", no abstractions for one use case. If you split one file into N similar files, stop and check whether N=1 with a few extra lines is clearer.
Surgical changes only. Touch what the task requires; nothing else. Don't "improve" adjacent code, comments, or formatting. Match existing style.
Dependency hierarchy — exhaust simpler first: vanilla language → standard library → external (only when stdlib can't do it in <100 lines). External deps must be maintained, lightweight, widely adopted. Exception: always use vetted libraries for security-critical code.
Lightweight over complex. Vanilla over frameworks. Express over NestJS, Flask over Django, plain JS over React if the project doesn't genuinely need it. Open-source only, no vendor lock-in.
Responsive web UI is mandatory. User builds web apps consumed on phones — fluid layouts, viewport meta, no horizontal scroll. Test in DevTools device emulation before claiming a UI task is done.
POC first. Validate logic with ~15min proof-of-concept before designing properly. Never ship the POC — rewrite it.
For full standards (testing, twelve-factor, deployment), see .claude/remember/AGENT_RULES.md.
@.claude/remember/MEMORY.md
Consult when building something new or adding a feature — a standards guide, not hot context like MEMORY.md above: @.claude/remember/AGENT_RULES.md