Imported from ninehills/skills (hunt/SKILL.md). Install upstream with npx skills add ninehills/skills --skill hunt. Copyright stays with the author.

Hunt: Diagnose Before You Fix

Prefix your first line with 🥷 inline, not as its own paragraph.

Update check (non-blocking). Before starting, run bash ../../scripts/check-update.sh once; if it prints a line, relay it to the user, then continue. It runs at most once a day, only reads a public version file, sends no data, and fails silently.

A patch applied to a symptom creates a new bug somewhere else.

Outcome Contract

• Outcome: the root cause is identified before any fix is applied.
• Done when: one sentence explains the cause, every observed symptom fits it, and the fix or handoff is verified against a reproducible check.
• Evidence: source trace, repro command or UI path, logs or state, targeted test/build output, and runtime evidence for UI or native defects.
• Output: root cause, fix or handoff, verification result, and any unswept sibling risks.

Do not touch code until you can state the root cause in one sentence:

"I believe the root cause is [X] because [evidence]."

Name a specific file, function, line, or condition. "A state management issue" is not testable. "Stale cache in useUser at src/hooks/user.ts:42 because the dependency array is missing userId" is testable. If you cannot be that specific, you do not have a hypothesis yet.

Diagnosis Signals

Good progress: a log line matches the hypothesis, you can predict the next error before running it, you understand the propagation path from root cause to symptom, you can write a test that fails on the old code. At each of these signals, find one more independent piece of evidence before committing.

Hypothesis quality gate: before acting on a hypothesis, list all observable symptoms (not just the one the user reported first). The hypothesis must explain every symptom; if it only covers some, it is a symptom-level guess, not a root cause. For timing-dependent issues (flicker, intermittent failure, race condition), reproduce reliably before diagnosing.

Rationalization warning: "I'll just try this" means no hypothesis, write it first. "I'm confident" means run an instrument that proves it. "Probably the same issue" means re-read the execution path from scratch. "It works on my machine" means enumerate every env difference before dismissing. "One more restart" means read the last error verbatim; never restart more than twice without new evidence.

Durable Context Preflight

See rules/durable-context.md for when to read durable context, the read-order budget, and the memory-type mapping.

For /hunt, diagnostic constraints are decision, preference, and principle entries; pattern and learning can seed hypotheses. Current code, logs, repro steps, tests, environment versions, and remote state override memory. Durable context is hypothesis fuel only. It never replaces a fresh root-cause sentence, a reproducible symptom list, or evidence from the current state.

Hard Rules

• Same symptom after a fix is a hard stop; so is "let me just try this." Both mean the hypothesis is unfinished. Re-read the execution path from scratch before touching code again.
• After three failed hypotheses, stop. Use the Handoff format below to surface what was checked, what was ruled out, and what is unknown. Ask how to proceed.
• Verify before claiming. Never state versions, function names, or file locations from memory. Run sw_vers / node --version / grep first. No results = re-examine the path.
• External tool failure: diagnose before switching. When an MCP tool or API fails, determine why first (server running? API key valid? Config correct?) before trying an alternative.
• System/tooling symptoms need a lower-layer baseline. Before blaming the visible app, generated file, or top-level feature, measure the raw lower layer first: OS capture versus post-processing, runtime service versus UI, compiler/toolchain versus test assertion, network/API versus client handling. Retire hypotheses that the baseline disproves instead of circling them.
• Pay attention to deflection. When someone says "that part doesn't matter," treat it as a signal. The area someone avoids examining is often where the problem lives.
• Visual/rendering bugs: static analysis first. Trace paint layers, stacking contexts, and layer order in DevTools before adding console.log or visual debug overlays. Logs cannot capture what the compositor does. Only add instrumentation after static analysis fails.
• Behavioral / lifecycle / async bugs: instrument first, not after failure. Window lifecycle, event delivery, navigation, focus, timer, state-machine, and async-ordering bugs almost never yield to static reading alone. Do not wait for a failed fix to add logs. The moment your hypothesis involves "this callback fires before/after that one", "this state should be X when Y runs", or "this object should still be alive here", add the log immediately as part of forming the hypothesis, before writing any fix. A hypothesis without runtime evidence is a guess; two guesses in a row is the hard-stop signal. Distinguish from visual-rendering bugs (compositor behavior needs DevTools, not logs) and pure-logic bugs (wrong formula, off-by-one) where static analysis is sufficient.
• Tuning magic numbers past round three: stop, unify. When a spacing / sizing / threshold value has been adjusted three times and still looks wrong, the bug is structural, not numeric. Replace the N independent values with one named token (Spacing.s4, --gap-content, etc.) and verify the asymmetry was hiding a missing constraint. Asymmetry that survives tuning is structural; more tuning will not converge.
• Fix the cause, not the symptom. If the fix touches more than 5 files, pause and confirm scope with the user.

Fix Scope Discipline

If the bug genuinely needs a refactor first (e.g. the cause cannot be addressed without changing a shared interface), pause, name the refactor explicitly, and ask. Do not silently bundle it. A bug fix that grew into a refactor is a separate PR.

Bisect Mode

Activate when: "以前是好的", "之前是好的", "used to work", "上一次提交还是对的", "broke after update", or the user remembers a specific good commit or version.

0. Protect the user's worktree first: run git status --short --branch -uall. If modified, staged, or untracked files exist, do not bisect in the current checkout. Create a temporary detached worktree from the same HEAD, run bisect there, then git bisect reset and remove the temporary worktree when done. If a temporary worktree is impossible, stop and ask for explicit cleanup/stash approval.
1. Find candidate good tag: git tag --sort=-version:refname | head -10 or ask the user for the last known-good commit. 1b. If the last-good version is only one or a few releases back, git diff <last-good-tag>..HEAD -- <suspect path> and read the delta directly first. The regression is usually visible in that diff, and reading it costs far less than driving a full bisect. Fall through to bisect only when the diff is too large or the culprit is not obvious.
2. Define a non-interactive pass/fail test command before starting bisect. Bisect is worthless without a reproducible check.
3. Run: git bisect start && git bisect bad HEAD && git bisect good <tag-or-hash>
4. At each step bisect checks out a commit. Run the test command. Mark: git bisect good or git bisect bad.
5. Let bisect drive. Do not jump ahead or skip commits unless explicitly asked.
6. When bisect names the culprit commit, read only that diff. Identify the specific line that introduced the regression.
7. Run git bisect reset when done.

Read large files once and reference from notes rather than re-reading at each bisect step.

Repeated Regression / Screenshot Reference Mode

Activate when the user says the same issue is still wrong after a fix, provides a "good" screenshot/version/file, or describes a visual result as previously correct.

Treat the reference as evidence, not decoration:

1. List every reported and visible symptom, preserving the user's concrete words where useful ("still slow", "not clear", "尖刺", "先显示上一个内容").
2. Identify the reference oracle: last-good commit/tag, old build, fixture, screenshot, downloaded artifact, or expected state from the user's description.
3. Define the pass/fail check before editing. For visual bugs, this may be a narrow screenshot checklist plus the command that renders the view; for behavioral bugs, prefer an automated regression test or deterministic repro.
4. Compare current vs. reference and name the exact delta. Do not generalize a visual defect into "style polish" when the evidence points to a broken render, race, font pipeline, or state path.
5. If the same symptom remains after one attempted fix, stop and rebuild the hypothesis from the evidence. Do not stack more patches onto a disproven explanation.

If the issue is purely subjective UI taste, route to /design. If it is rendering, state, timing, build output, font generation, or a regression from a known-good version, stay in /hunt.

Scope Blast Mode

Activate after fixing a root-cause pattern, before declaring the bug done; also when the user says "举一反三", "举一反三深入看看", or "其他地方有没有同样问题". The same shape often hides in N other places; one local fix that ignores the blast leaves N - 1 bugs in the tree.

1. Extract the pattern signature: the specific function name, regex, API call, CSS selector, lock acquisition, validation skip, or input boundary that produced the bug.
2. grep -rn <pattern> across the repo (exclude generated dirs, build output, vendored deps). For class-of-bug patterns (e.g. "any handler missing the lock"), grep for the surrounding shape, not just the literal text.
3. List every match. For each one, answer in writing: same bug here? Pick fix / leave (explain why it is safe) / unsure (ask the user). Do not silently skip a match.
4. Do not claim "fixed" until the blast report is in the Outcome block.

Common triggers:

• Visual bug fixed on one page: check every other page using the same component, layout, or media-query breakpoint.
• One race fixed in one handler: check every handler acquiring the same lock or touching the same shared state.
• One validation skip patched at one entry point: check every entry point that reaches the same downstream sink.
• One regex / parser fix for one input shape: check every caller of the same regex / parser.

If the blast surfaces unrelated bugs, list them but do not fix in this PR unless the user agrees; scope creep is its own anti-pattern.

Confirm or Discard

The instrument-first rule lives in Hard Rules (behavioral/async bugs) above; this is what to do with its result. Run the one probe that would fail if the hypothesis were wrong, then read it. If the evidence contradicts the hypothesis, discard it completely and re-orient on what the probe just showed. Do not stack a fix onto a disproven hypothesis, and do not keep one just because the code "looks like" the cause.

Runtime Evidence Ladder

Use this ladder before claiming a bug is fixed:

1. Source trace: name the exact function, state transition, file, line, or condition that can produce the symptom.
2. Deterministic repro: run or write the smallest command, fixture, UI path, or scenario that produces it.
3. Logs/state/cache: inspect the runtime state that proves the path was reached, including queues, DB rows, caches, temp files, generated outputs, or external tool logs.
4. Build/test: run the narrow test or build that exercises the fix.
5. Real runtime check: for UI, native app, browser, rendering, or visual bugs, open the app/page/artifact and verify the visible result with a screenshot or concrete checklist.

Compile-only is not enough for UI, native-app, visual, rendering, or generated-artifact bugs. If the runtime check is impossible in the environment, say why and hand off the exact screen, command, or artifact to verify.

For recurring classes of failures, load references/failure-patterns.md before adding a second fix.

Native App Freeze Mode

Activate when a desktop or mobile native app reports beachball, not responding, tab-switch freeze, first-open lag, idle wake stall, overlay lockup, or a screenshot shows a frozen app.

Evidence to collect before changing code:

1. Exact user path and version: first launch versus warm launch, the tab or window transition, idle duration, permissions, display count, and any setting that makes the freeze disappear.
2. Runtime capture while frozen: sample <process>, recent app logs, CPU and memory footprint, thread count, and whether the main thread is blocked, spinning, or allocating.
3. First-frame surface: view body work, first .task, synchronous icon or metadata lookup, filesystem scans, URL parent walks, notification callbacks, and app/window wake handlers.
4. Blast search after the fix: grep the same API shape across the repo, especially path parent walks, synchronous icon loading, metadata reads in render paths, and callbacks that run on the main thread.

Common native freeze traps:

• Launch, terminate, permission, audio, display, or workspace notifications doing path walks, icon lookup, filesystem scans, or process enumeration on the main thread.
• First paint hydrating a full app list, directory tree, media thumbnail set, or system status table before showing an interactive shell.
• An input-lock or full-screen overlay without a guaranteed teardown path for Escape, app deactivation, permission denial, process termination, and window close.
• Timer or sampler work that survives hidden windows, long idle periods, sleep/wake, or app reactivation.

Compile-only and source-only checks are insufficient for this mode. The outcome must include the runtime capture, the root-cause frame or state transition, the focused regression guard, and any sibling matches that were fixed or explicitly left safe.

Targeted Logging

Use logs as a scalpel, not as noise. Before adding a log, write the question it answers:

"If this log prints X before Y, hypothesis A is still possible; if it does not, hypothesis A is wrong."

Load references/logging-techniques.md for the full logging playbook: binary-search instrumentation, discriminating log content, boundary-first placement, timing bug logging, and removal discipline.

Quick rules:

1. Place the first log at the midpoint of the execution path, not at the symptom. Binary search from there.
2. Log discriminating facts only: sequence number, input key, branch taken, old/new state, error code.
3. Remove temporary logs before finishing. Gate persistent diagnostics behind the project's debug flag.

If adding logs changes the behavior, treat that as evidence of a timing, lifecycle, or concurrency problem.

Gotchas

Outcome

Success Format

Root cause:        [what was wrong, file:line]
Fix:               [what changed, file:line]
Confirmed:         [evidence or test that proves the fix]
Tests:             [pass/fail count, regression test location]
Regression guard:  [test file:line] or [none, reason]

Status: resolved, resolved with caveats (state them), or blocked (state what is unknown).

Regression guard rule: for any bug that recurred or was previously "fixed", the fix is not done until:

1. A regression test exists that fails on the unfixed code and passes on the fixed code.
2. The test lives in the project's test suite, not a temporary file.
3. The commit message states why the bug recurred and why this fix prevents it.

Handoff Format (after 3 failed hypotheses)

Symptom:
[Original error description, one sentence]

Hypotheses Tested:
1. [Hypothesis 1] → [Test method] → [Result: ruled out because...]
2. [Hypothesis 2] → [Test method] → [Result: ruled out because...]
3. [Hypothesis 3] → [Test method] → [Result: ruled out because...]

Evidence Collected:
- [Log snippets / stack traces / file content]
- [Reproduction steps]
- [Environment info: versions, config, runtime]

Ruled Out:
- [Root causes that have been eliminated]

Unknowns:
- [What is still unclear]
- [What information is missing]

Suggested Next Steps:
1. [Next investigation direction]
2. [External tools or permissions that may be needed]
3. [Additional context the user should provide]

Status: blocked

Rendering Bug Mode

Activate when: "PDF looks wrong", "page break issue", "font not rendering", broken PDF output, or print layout wrong.

Load references/rendering-debug.md for the full diagnosis checklist (WeasyPrint quirks, font loading, page overflow, browser print CSS). Static analysis first, then reproduce if needed.

IME / Unicode Issues

For input method, character rendering, or text encoding bugs (IME state, cursor drift, emoji splitting, composition events), check references/ime-unicode.md first before forming a hypothesis.