Create a new versioned release with changelog. Bumps version in code, updates CHANGELOG.md, commits, tags, and pushes using the repository release flow. GitHub Actions creates the release and uses only the current changelog section as release notes. Use when the user says "release", "cut a release", "bump version", "new release".
Review changed code for correctness bugs — logic errors, nil/error handling, concurrency, resource leaks, edge cases, broken invariants. Reports findings first, ordered by severity, then fixes on request. Scoped to correctness, not quality or style cleanups. Use when the user says "code review", "review my changes", "any bugs", "find bugs", or asks whether a change is correct.
Systematic debugging that finds the root cause before changing any code — collect symptoms, trace the code path, check what changed, reproduce deterministically, form a testable hypothesis, then fix the cause (not the symptom) and prove it with a regression test. Use when the user reports a bug, a crash, a failing test, or says "debug", "why is this happening", "root cause".
Restructure existing code without changing its behavior — extract, rename, inline, split, deduplicate, reshape — with tests holding behavior fixed before and after. Scoped to structure only: no new features, no bug fixes, no behavior changes. Use when the user says "refactor", "clean up the structure", "extract this", "split this file", "untangle", or "restructure".
Review changed code for reuse, quality, and efficiency, then clean up what you find — duplicated logic that could reuse an existing helper, hacky or redundant patterns, needless work on hot paths. Quality cleanup only: not bugs (that's code-review) and not behavior-preserving restructuring on request (that's refactor). Use when the user says "simplify", "clean this up", "is there a simpler way", or "tidy the diff".
Test-driven development — write a failing test that names the behavior, watch it fail, implement the minimum to make it pass, then refactor with tests green. Works for new features, bug fixes, and behavior changes. Use when the user says "tdd", "test-first", "write tests first", or wants a change built test-first.
Regression test a San feature by name. Looks for a feature doc in docs/reference/ or docs/packages/2-feature/, runs automated Go tests and interactive tmux tests, then produces a pass/fail report. Use this skill when the user says "qa", "regression test", "test feature X", "verify feature", or references a feature name like "agent", "cli-startup", "loop".
Review changed code for reuse, quality, and efficiency, then fix any issues found.