메뉴
Analyze dogfood results and classify failures by root cause
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
Suggest and apply a semver version bump based on commits since the last tag
Push current branch to remote origin
Scaffold a new Sharpy.Stdlib module (spy-sourced or handwritten C#) with all required files, conventions, docs, and tests
Implement a plan with a coordinated agent team
Run compiler or cross-language benchmarks and compare results
Regenerate C# snapshot tests and spy stdlib after intentional codegen changes
| name | dogfood-analyze |
| description | Analyze dogfood results and classify failures by root cause |
| argument-hint | [directory_name] |
Analyze dogfood results in dogfood_output/ and classify each failure into root cause categories (C1-C5).
Use the dogfood-analyst agent via the Task tool to perform the analysis:
Task tool:
subagent_type: dogfood-analyst
prompt: (see below)
If $ARGUMENTS is empty (no specific directory given), analyze all results:
Analyze all dogfood results in dogfood_output/. Read SUMMARY.md and runs.json first, then investigate each issue and skip directory. Classify every failure using categories C1-C5 and produce a structured report.
If $ARGUMENTS specifies a directory name, analyze only that item:
Analyze a single dogfood result. Look for the directory matching "$ARGUMENTS" under dogfood_output/issues/ or dogfood_output/skips/. Read its metadata.json, error.txt or skip_reason.txt, and source files. Classify using categories C1-C5 and produce a focused report for this single item.
Present the agent's report directly to the user.