Agent skill
tools-bug-scan
Run the repository-native bug scanner (`pnpm bug-scan` / `pnpm bug-scan:changed`), summarize findings by severity and rule, and propose concrete next fixes or targeted rescans.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/tools-bug-scan
SKILL.md
Bug Scan Tool
Use this skill when the user wants a fast, deterministic bug-pattern sweep using the in-repo scanner.
Commands
- Full/targeted scan:
pnpm bug-scan [path ...]
- Changed-files scan:
pnpm bug-scan:changed
- Machine-readable output:
pnpm bug-scan -- --format=json [path ...]pnpm bug-scan -- --changed --format=jsonpnpm bug-scan -- --changed --format=json --fail-on=none --business-scope=<BIZ> --idea-artifact=docs/plans/<slug>/bug-scan-findings.user.json
- Rule discovery and tuning:
pnpm bug-scan:rulespnpm bug-scan -- --changed --fail-on=criticalpnpm bug-scan -- --only-rules=<ruleA,ruleB>pnpm bug-scan -- --skip-rules=<ruleA,ruleB>pnpm bug-scan -- --write-baseline=tools/bug-scan-baseline.json --fail-on=nonepnpm bug-scan -- --baseline=tools/bug-scan-baseline.json
Default Workflow
- Choose scope:
- If user asks for changed work, run
pnpm bug-scan:changed. - If user provides files/directories, pass them to
pnpm bug-scan. - If unspecified, prefer changed-files first to keep signal high.
- If user asks for changed work, run
- Run scanner once.
- If findings exist, summarize by:
- severity counts (
critical,warning) - rule IDs
- top file locations (
path:line:column)
- severity counts (
- Recommend a concrete next action set:
- immediate fixes for
critical - optional cleanup queue for
warning - exact rerun command after fixes
- immediate fixes for
- If the scan is part of post-build reflection, write the idea artifact into the active plan directory so
startup-loop:generate-process-improvementscan ingest findings as idea candidates.
Output Contract
Always report:
- Scan scope used.
- Finding totals by severity.
- Top findings with file references.
- Exit status implication:
0= no findings1= findings present
Guardrails
- Do not install external scanners or mutate hooks/shell profiles in this skill.
- Do not fail the task if scanner reports findings; report and propose fix order.
- If scanner command fails for runtime/tooling reasons, report the error and suggest the minimal unblocking command.
Escalation
Escalate to planning (lp-do-fact-find / lp-do-build) when:
- the same rule repeatedly appears across many files,
- findings indicate systemic pattern debt,
- user asks to convert findings into an implementation plan.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?