Skill Claude CodeCodex
How to build an AI agent that actually works — start with a decision-map that forces a singular outcome (or splits into multiple agents), design one in-session, then optionally fan out parked agents in parallel after confirm. Locks Task · Context · Memory · Eval. Use when designing, building, reviewing, or debugging…
Skill Claude CodeCodex
Opt-in only. Write, resume, or fetch/hydrate a named Cursor ↔ Claude Code handoff under .scratch/handoffs/. File first; else hydrate from the other tool's local session store (confirm required). Diff check warns if shipped ≠ tree — never blocks. Switch-only, not live sync. Do not run proactively.
Skill Claude CodeCodex
Verify a frontend change ACTUALLY rendered before calling it done. Inspect the served HTML + compiled CSS on the live dev server — never trust source edits or a normal browser refresh. Catches the dev-server staleness trap (Fast Refresh silently drops edits) that makes every "fix" look identical to the last. Use after…