메뉴
Audit and rank scip-query cleanup signals without editing code. Use for health reports, de-bloat reports, recent AI residue audits, score-framed cleanup queues, confirming raw findings, or preparing a cleanup plan.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | scip-cleanup-audit |
| description | Audit and rank scip-query cleanup signals without editing code. Use for health reports, de-bloat reports, recent AI residue audits, score-framed cleanup queues, confirming raw findings, or preparing a cleanup plan. |
| commands | [{"template":"scip-query health --json","when":"Establish evidence: composite score and prioritized action list."},{"template":"scip-query cleanup-plan --verify --json","when":"Sweep signals: compiler-verified deletion plan, batched."},{"template":"scip-query duplicate-bodies --json --full","when":"Sweep signals: exact duplicate small-body candidates."},{"template":"scip-query recent-duplicates --json --full","when":"Sweep signals: recent code that re-implements established code."},{"template":"scip-query incomplete-migration --json --full","when":"Sweep signals: partially-completed extractions."},{"template":"scip-query doc-drift --json --full","when":"Sweep signals: stale docs referencing changed code."}] |
Use this skill to turn raw scip-query signals into a confirmed cleanup queue. Do not edit application code in this skill.
Load shared mechanics from ../_shared/SKILL.md.
| Command | Purpose | When |
|---|---|---|
scip-query health --json | Composite codebase health report with prioritized action list | Establish evidence: composite score and prioritized action list. |
scip-query cleanup-plan --verify --json | Ordered, batched deletion plan: graph-fact dead code plus the cascade candidates it unlocks | Sweep signals: compiler-verified deletion plan, batched. |
scip-query duplicate-bodies --json --full | Find exact duplicate small-body candidates across files | Sweep signals: exact duplicate small-body candidates. |
scip-query recent-duplicates --json --full | Directional duplicate candidates: recent code that re-implements established callable, React, or Vue code | Sweep signals: recent code that re-implements established code. |
scip-query incomplete-migration --json --full | Partially-completed extraction candidates: new helpers in the diff wired into some sites while similar un-migrated sites remain | Sweep signals: partially-completed extractions. |
scip-query doc-drift --json --full | Stale-doc candidates: code the doc references or co-changed with kept changing after the doc stopped | Sweep signals: stale docs referencing changed code. |
Use this shortlist first. Open ../_shared/SKILL.md only when it is insufficient.
scip-query doctor
scip-query status --capabilities
scip-query health --json
scip-query capabilities --json
scip-query config-validate --json
This step is complete only when unavailable capabilities are recorded as unavailable.
Run every relevant class, or record why it is unavailable:
scip-query cleanup-plan --verify --json
scip-query duplicate-bodies --json --full
scip-query recent-duplicates --json --full
scip-query incomplete-migration --json --full
scip-query unused-params --json --full
scip-query passthrough-candidates --json --full
scip-query dead --json --full
scip-query isolated --json --full
scip-query cycles
scip-query co-change --json --full
scip-query doc-drift --json --full
For frontend repos, add the React or Vue duplicate, hook/composable, and large-component/view commands.
Optional deep-dive (near-zero precision on codebases with intentional layering/ambient types — run only when the sweep above is exhausted, and treat every hit as a lead to confirm, not a finding):
scip-query stale-abstractions --json --full
scip-query wrapper-candidates --json --full
For each high-priority candidate, inspect source and graph evidence:
scip-query code <symbol-or-file>
scip-query refs <symbol>
scip-query fan-in <symbol>
scip-query fan-out <symbol-or-file>
scip-query affected <symbol> --json
scip-query change-surface <file> --json --full
scip-query similar <symbolA> <symbolB> --plan
scip-query co-change <file> --json --full
Classify every candidate as confirmed fix target, intentional design, false positive, or blocked.
A deletion is this skill's scrutiny-ending verdict, so a confirmed fix target that deletes code must survive refutation first. Run the known blind-spot checks and record them on the entry: rg for the symbol name as a string (dynamic dispatch, config keys, serialized references, CLI/doc text), and the cross-package barrel re-export gap (the one shape dead self-labels unconfirmed). Record refutation: survived — <checks run> on the entry, or reclassify it; the note stays either way.
Health score: N/100
Classification counts: <n> confirmed / <i> intentional / <fp> false positive / <b> blocked (must cover every collected signal)
Confirmed items:
- [priority] finding - evidence - first safe action
Unconfirmed signals:
- signal - evidence still needed
Unavailable or blocked checks:
- check - reason
Recommended first cleanup batch:
- batch - why safe now
The audit is complete only when each collected signal is classified and the next action is visible.
Assess API impact with scip-query evidence. Use before changing public exports, module boundaries, schemas, routes, CLI commands, config fields, generated artifacts, signatures, docs-backed behavior, or consumer migrations.
Audit output-facing status claims for evidence with scip-query. Use to classify whether an "available", "verified", "safe", "PASS", or "complete" status word is derived from a real check, hedged as a candidate, or merely asserted without being probed.
Plan code changes with scip-query evidence and testable design. Use for non-trivial implementation, refactor, migration, API, or bug-fix plans before editing code; require contextual definitions, cited premises, reuse audit, test seams, counterexample attacks, and a derived verdict.
Debug bugs and regressions with scip-query evidence. Use for failing behavior, wrong data flow, confusing runtime paths, broken tests, root-cause analysis, reproduction, tracing, or minimal fixes.
Explore codebases with scip-query evidence. Use to explain how a system, feature, module, call path, dependency graph, data flow, architecture, or change risk works before answering or editing.
Audit whether implementations are real with scip-query. Use for suspected faked or half-implemented features, decorative checkers or verifiers that never fail, dead code paths hidden behind graceful fallbacks, metrics that may be lying, or a "does any of this actually work" interrogation of a system.