メニュー
X-ray a codebase before reading it. Runs 5 git diagnostics — churn hotspots, bus factor, bug clusters, commit velocity, and revert frequency — then delivers a brief risk report. Use when the user says "xray", "scan this repo", or invokes /xray.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Compile the Obsidian wiki/ layer from raw/ atomic notes. Diffs which notes have not yet been cited in any wiki sources[], routes each into a new or existing concept article, writes provenance-carrying markdown, and lints the result against wiki_graph(). Use when the user says "compile", "compile the wiki", "rebuild concepts", "compile-wiki", or invokes /compile-wiki. Requires the obsidian MCP (raw_graph, wiki_graph, read, write).
Set up nia.md at the current project root. Analyzes dependency manifests, filters to the libraries worth binding, resolves them against Nia, and writes nia.md using a fixed template. Then asks if the user wants to add docs or other repos. Use when the user says "set up sources", "create nia.md", "bind nia to this project", or invokes /sources.
Pull context from a previous entire-tracked session into the current chat. Lists sessions in the current repo (or any path passed as arg), lets the user pick when more than one matches, writes a 6-section compaction summary to entire_context_<id>.md in cwd, and feeds the same summary into chat. Use when the user says "bring context", "import context", "pull from session", or invokes /bring-context.
Analyze GLPI tickets at depth — pick the right MCP tool, pull full ticket content in one shot, and delegate large-result-set analysis to an Explore agent instead of writing Python. Use when the user asks about GLPI tickets, requests patterns/themes across tickets, asks "what's happening with X", "find anything related to Y", "tickets this week/month", or invokes /glpi.
Safely commit and push changes. Audits .gitignore for sensitive files, stages with git add ., writes a conventional commit (what + why), creates a GitHub repo via gh if none exists, and pushes. Use when the user says "push", "ship it", "commit and push", or invokes /git-push.
Fill vendor skill gaps when a built-in or custom skill reaches for the wrong tool or misses a vendor-specific workflow. Use proactively when a skill starts using the wrong integration, when the user asks to use a specific vendor workflow, or when invoked as /vendor-sklls-gap.
| name | xray |
| description | X-ray a codebase before reading it. Runs 5 git diagnostics — churn hotspots, bus factor, bug clusters, commit velocity, and revert frequency — then delivers a brief risk report. Use when the user says "xray", "scan this repo", or invokes /xray. |
Run all 5 diagnostics, then deliver a single concise report. No preamble, no explaining what you're about to do — just run the commands and present findings.
Run all 5 in parallel:
git log --format=format: --name-only --since="1 year ago" | sort | uniq -c | sort -nr | head -20
git shortlog -sn --no-merges
git log -i -E --grep="fix|bug|broken" --name-only --format='' | sort | uniq -c | sort -nr | head -20
git log --format='%ad' --date=format:'%Y-%m' | sort | uniq -c
git log --oneline --since="1 year ago" | grep -iE 'revert|hotfix|emergency|rollback'
After running the commands, deliver a short report with these sections:
Hottest files — top 5 highest-churn files. Flag any that also appear in the bug cluster list (these are the highest-risk files).
Bus factor — how concentrated is ownership? Call out if one person is 60%+ of commits. Note if top contributors appear inactive.
Bug magnets — files that keep breaking. Cross-reference with churn — files on both lists are the biggest risk.
Velocity — is the project accelerating, decelerating, or steady? Any gaps that suggest people leaving or momentum dying?
Stability — how many reverts/hotfixes in the last year? A few is normal. Every couple weeks means deploy trust is low.
End with a one-line verdict: is this codebase healthy, manageable, or on fire?