菜单
Hypothesis-driven defect isolation — stack-trace forensics, breakpoint strategy, state inspection, and root-cause confirmation via minimal repro. Use when a defect surfaces (test failure, crash, exception, wrong output, intermittent flake) and the cause is not immediately obvious from the change set.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Verbalized Sampling (VS) protocol for intent exploration before planning, mode-aware. Default `exhaustive` runs full VS; `collaborative` runs tip-sharing dialogue; `adversarial` walks the design tree one fork at a time. Auto-detects from phrasing ("help me refine" → collaborative, "poke holes" → adversarial); override via `/askme adversarial|collaborative|exhaustive`. Use for ambiguous tasks or maximum clarifying questions before committing.
Tidy-operations dispatcher. Invoke on "tidy", "clean up", "tidy this file/memory/workspace/git/docs", or when active context (current file, diff, stack, memory directory) has structural rot to resolve before touching behavior. Detects target domain from context and routes to the sibling skill. Requires explicit target or clear active-context signal — do not invoke speculatively.
Read-only codebase exploration: discovery, structural reading, and emission of architecture/pattern/tooling/dependency summaries. Use to understand existing code, map files, trace function flow, locate symbols, or build pre-implementation context. Uses 1/3/5 Explore-agent escalation scaled to scope. Trigger on "explore", "find where X is", "how does X work in the code", "map the codebase", "what files handle Y", or any architecture/pattern/tooling/dependency context request on a local repo — even without naming /explore.
Surface deepening refactors that turn shallow modules into deep ones, informed by `CONTEXT.md` and `docs/adr/`. Use when the user asks to improve architecture, find refactor candidates, raise testability, or make a codebase more agent-navigable. Skip for single localized fixes.
Surface in-task-collaboration protocols when the user describes an AI workflow informally — URL-as-entity-reference, PR-comment threads as session memory. Trigger when the user names entities by colloquial label instead of stable URL, asks "how should I structure this for Claude", or describes a multi-step Claude workflow without a durable handle. Apply reactively, not as a checklist.
Code search, analysis, and refactoring using ast-grep (sg). Use for AST-based code modifications, structural search, and linting.
| name | debug |
| description | Hypothesis-driven defect isolation — stack-trace forensics, breakpoint strategy, state inspection, and root-cause confirmation via minimal repro. Use when a defect surfaces (test failure, crash, exception, wrong output, intermittent flake) and the cause is not immediately obvious from the change set. |
A bug is a falsified assumption. Find the assumption, falsify it deliberately, observe the divergence, narrow until one line owns the lie. No speculation, no shotgun edits, no "fix and rerun" guessing.
Apply: test fails and cause unclear; production stack trace; intermittent / flaky behavior; wrong output without crash; regression after known commit window; heisenbug.
NOT apply: performance regression with correct outputs; security defect; symptom obvious from one-line read; architectural confusion.
git bisect for regressions.caused by chain encodes why; the stack encodes where.<inlined> markers signal source-line-to-instruction map is approximate.| Family | Live debugger | Postmortem / record | Remote attach |
|---|---|---|---|
| Systems (C/C++/Rust) | gdb, lldb, rust-gdb, rust-lldb | coredumpctl + gdb core, rr record/replay | gdb -p <pid> / lldb -p <pid> |
| Python | pdb, ipdb, pdbpp, breakpoint() | faulthandler, py-spy dump, traceback module | debugpy --listen |
| Go | dlv debug, dlv test, dlv attach <pid> | runtime/pprof, GOTRACEBACK=crash | dlv connect <addr> |
| Java/Kotlin | IntelliJ debugger, jdb | hs_err logs, JFR, heap dump (jmap) | JDWP -agentlib:jdwp=... |
| JavaScript/TypeScript | node --inspect, Chrome DevTools | --report-uncaught-exception reports | --inspect=0.0.0.0:9229 |
| OCaml | ocamldebug, Printexc.record_backtrace true | core file + ocaml-gdb, memtrace | ocamldebug -s <socket> |
Use procs (not ps) for PID. Use bat -P -p -n (not cat) for trace files. Use git grep -n -C 3 'pattern' (not grep) for callsites.
Before hypothesizing a fix, reason through the failure — SHORT-form KEYWORDS for trace notes, observe the symptoms, trace the execution path, break down where actual behavior diverges from expected, critically review each candidate cause, validate each hypothesis against the evidence. The root cause is the smallest explanation that accounts for all observed symptoms. For numeric calculation (timing math, bound arithmetic, off-by-N analysis), invoke fend per the baseline rule; never self-calculate. Causal reasoning and trace interpretation are in-head — they are not arithmetic.