autopilot

This directory ships **three methodology agents** bundled with autopilot. They carry autopilot's discipline axis into agent-level execution so dispatched reviews, debugs, and plans follow the same Three Red Lines as the skill layer.

Use when encountering any bug, service outage, test failure, flaky test, or unexpected runtime behavior — applies autopilot's evidence-first methodology (gather → narrow → hypothesize → verify → propose-fix). Never guesses from memory. Triggers PUA mode after 2+ failed attempts. Read-only diagnostic — produces fix proposals, does not apply patches.

Use when breaking fuzzy requirements into parallelizable subtasks, decomposing L-size work, or producing structured Task Breakdowns — applies autopilot's six-element Task Prompt contract (goal / scope / input / output / acceptance / boundaries). Read-only — does not write code or apply edits. Dispatched by autopilot:dev-flow on L-size tasks and by autopilot:think-tank for structured decomposition.

Use when performing pre-commit / pre-merge review, security audit, or plan critique — applies autopilot's Three Red Lines (closure / fact-driven / exhaustiveness). Every finding cites file:line. Default-assumes everything is broken until verified. Dispatched primarily via autopilot:quality-pipeline.

Compare two implementations and find every difference. Use when: "compare X with Y", "check X against Y", "verify X matches Y", "flag anything missing", "比對 X 和 Y", "檢查有沒有漏掉", "驗證是否一致", feature parity review between old and new systems, spec vs implementation verification, cross-system completeness check. Not for debugging a single discrepancy or writing comparison tests.

Pre-code Socratic design exploration — when the options don't exist yet. Use when: "I haven't figured out the approach", "help me think through X", "what should this even be / 還沒想清楚", "explore how to do X before we build", "我不確定要怎麼做", "尚未成形 / 沒有方向". DISCOVERS the options, surfaces 2-3 genuinely different approaches, and GATES implementation until you approve a design. Not for: deciding between options that ALREADY exist on the table (→ think-tank), external best-practice research (→ survey), an approach you've already chosen (→ references/plan-template.md to author the plan, then dev-flow), or a quick known fix (→ dev-flow).

Full delegation — you own the goal end-to-end, user just wants results. Triggers: "CEO mode", "get it done", "you decide", "handle everything", "I trust you", "take over", "full authority", "just do it", "搞定 X", "幫我處理", "全權處理", "你決定". Not for: research-only (→ survey), participatory planning (→ dev-flow), or parallel task dispatch.

Evidence-first debugging for correctness issues. Invoke when diagnosing bugs, crashes, logic errors, data corruption, connectivity problems, intermittent failures (incl. flaky tests with environment divergence), or 'works on my machine' issues. For performance issues (slow / high latency / CPU / memory), use the profiling skill instead. For perf regressions with known change attribution, profiling is still primary — measure first, don't guess from the deploy diff.

Start here before writing any code — sizes task (S/L/H/Fix), sets up branch and session rules. Use when: "I'm starting on X", "quick fix for Y", "continuing from yesterday", "hotfix needed", "let's implement X", "skip to coding", "我要開始做 X", "快速修一下", "接續昨天的進度", resuming a feature branch, or any task that touches code. Not for: debugging (→ systematic-debugging), authoring a plan doc (→ references/plan-template.md), pre-code design exploration (→ brainstorm), or code review (→ quality-pipeline).

8 hooks across 3 events