Claude Skills by chrille0313

Run the inner agentic loop on one task. The implementer builds, a fresh-context reviewer judges, and findings cycle back until approval (max 3 rounds). Use when the user says "build X", "implement issue #N", "work on the next backlog item", or CI dispatches a labeled issue.

Commit working changes as small, reviewable conventional commits. Use when asked to commit, and at the end of any task before opening a PR.

Interactively plan a feature with the user, then land it as right-sized build-ready tasks: one spec handed straight to /build, or dependency-ordered backlog issues. Use when the user says "plan X", "break down X", "create issues for X", or brings an idea too large for a single /build task.

Run the outer agentic loop once. Triage the backlog, pick the most valuable ready item, shape its spec, and hand it to the inner loop. Use when the user asks "what's next", "groom the backlog", "run the PM", or on a schedule. Pass --dispatch to label the chosen issue for CI instead of building locally.

One-time template adaptation. Pick a language/framework, scaffold a minimal project, and fill in the command contract in CLAUDE.md that all agents verify against. Use right after creating a repo from this template, or to swap the stack later.