メニュー
Derive test regression plans from git history by finding commits lacking tests. Use when finding untested bug fixes, coverage gaps, or generating a test backlog.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
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。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/laurigates/claude-plugins --skill blueprint-derive-testsコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SOC 職業分類に基づく
SKILL.md を表示中
| created | "2026-02-22T00:00:00.000Z" |
| modified | "2026-06-18T00:00:00.000Z" |
| reviewed | "2026-06-10T00:00:00.000Z" |
| description | Derive test regression plans from git history by finding commits lacking tests. Use when finding untested bug fixes, coverage gaps, or generating a test backlog. |
| args | [--since DATE] [--quick] [--scope AREA] |
| argument-hint | --since 2024-06-01 for date range, --quick for last 50, --scope auth for specific area |
| allowed-tools | Read, Write, Glob, Grep, Bash, AskUserQuestion, Task |
| model | opus |
| name | blueprint-derive-tests |
Analyze git history to identify fix and feature commits lacking corresponding test changes, then generate a structured Test Regression Plan (TRP) document as a prioritized test backlog.
Use case: Systematically close test coverage gaps by mining commit history for bug fixes and features that shipped without regression tests.
| Use this skill when... | Use alternative when... |
|---|---|
| Bug fixes ship without regression tests | You need to run existing tests (/test:run) |
| Want a prioritized test backlog from history | Writing tests for a specific feature (manual TDD) |
| Onboarding a project and assessing test health | Checking current test coverage metrics |
| Need to find which fixes lack test coverage | Designing a test strategy from scratch (/test:architecture) |
git rev-parse --git-dirfind . -path '*/docs/blueprint/*' -maxdepth 3 -name 'manifest.json' -type fgit rev-list --count HEADfind . -maxdepth 3 \( -name 'vitest.config.*' -o -name 'jest.config.*' -o -name 'pytest.ini' -o -name 'pyproject.toml' -o -name 'Cargo.toml' -o -name 'go.mod' \) -type f -print -quitfind . -maxdepth 4 -type f \( -name '*.test.*' -o -name '*.spec.*' -o -name 'test_*' -o -name '*_test.*' \) -printgit log --format="%s" --max-count=10Parse these from $ARGUMENTS:
--quick: Fast scan (last 50 commits only)--since DATE: Analyze commits from specific date (e.g., --since 2024-06-01)--scope AREA: Filter to commits touching a specific area/scope (e.g., --scope auth)Default behavior without flags: Analyze last 200 commits.
For detailed templates, severity matrix, and test mapping rules, see REFERENCE.md.
Execute this test regression plan derivation workflow:
Check context values above:
/blueprint:init, then continueParse $ARGUMENTS for --quick, --since, and --scope:
--quick → scope = last 50 commits--since DATE → scope = commits from DATE to now--scope AREA → filter commits to those with scope matching AREA or touching files in AREA directoryStore scope parameters for git log commands in subsequent steps.
Scan for test framework and conventions:
*.test.ts, *.spec.ts (JS/TS)test_*.py, *_test.py (Python)*_test.rs, tests/ directory (Rust)*_test.go (Go)src/ → tests/, src/ → src/__tests__/)If no test framework detected → Warn user, continue with file-based detection only.
Run the helper. It owns the deterministic core: classifying fix:/feat:
commits from git log, detecting whether each commit carried an inline
test-file change (git show --name-only), and assigning a severity via the
fixed matrix (fix: + no inline test → CRITICAL; feat: + no inline test →
MEDIUM; any commit shipping a test → COVERED). Pass --limit to cap the scan
(default 200; pass --limit 50 for --quick):
bash "${CLAUDE_SKILL_DIR}/scripts/blueprint-derive-tests.sh" --home-dir "$HOME" --project-dir "$(pwd)"
Parse STATUS= and ISSUES: from the output. FIX_COMMITS/FEAT_COMMITS are
the classification counts; GAPS_CRITICAL/GAPS_MEDIUM/GAPS_TOTAL are the
coverage gaps; each coverage_gap issue carries SHA=, TYPE=, SEVERITY=,
and SUBJECT= for the TRP table. STATUS=ERROR means at least one CRITICAL
(untested fix) gap.
The script's inline-test detection and base severity (CRITICAL/MEDIUM) are the
deterministic floor. When you need the finer High/Low tiers, nearby-test-commit
softening, or the per-language source-to-test mapping, apply the modifiers from
REFERENCE.md and the mapping rules in
REFERENCE.md. For --since/--scope
filtering, narrow the git scope with the commands in
REFERENCE.md before reading the gap set.
TRPs live at the top level under docs/trps/ — not docs/blueprint/trps/. This matches the sibling derive-* skills' top-level layout. Never write TRPs under docs/blueprint/; that path is reserved for blueprint machinery.
mkdir -p docs/trpsid_registry.last_trp, increment by 1TRP-001regression-gaps-2024-q3)docs/trps/{slug}.md using template from REFERENCE.mdInclude in the document:
id, status: Active, scope, date_range, commits_analyzedIf Blueprint is initialized:
id_registry.last_trp with the new TRP numberid_registry.documents:
{
"TRP-NNN": {
"path": "docs/trps/{slug}.md",
"title": "{TRP title}",
"status": "Active",
"created": "{date}"
}
}
jq --arg now "$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
--arg sha "$(git rev-parse HEAD 2>/dev/null)" \
--argjson analyzed "{commits_analyzed}" \
--argjson gaps "{gaps_found}" \
'.task_registry["derive-tests"].last_completed_at = $now |
.task_registry["derive-tests"].last_result = "success" |
.task_registry["derive-tests"].stats.runs_total = ((.task_registry["derive-tests"].stats.runs_total // 0) + 1) |
.task_registry["derive-tests"].stats.items_processed = $analyzed |
.task_registry["derive-tests"].stats.items_created = $gaps |
.task_registry["derive-tests"].context.commits_analyzed_up_to = $sha' \
docs/blueprint/manifest.json > tmp.json && mv tmp.json docs/blueprint/manifest.json
Print summary:
Test Regression Plan Generated!
**Analysis Summary**
- Commits analyzed: {N} ({date_range})
- Fix commits found: {N}
- Feature commits found: {N}
**Coverage Gaps Found**
- Critical: {N} (fix commits with no tests at all)
- High: {N} (fix commits with stale test files)
- Medium: {N} (feature commits missing tests)
- Low: {N} (feature commits with nearby tests)
**Document**: docs/trps/{slug}.md (TRP-{NNN})
**Top Priority Gaps**
1. {commit subject} — {severity} — {affected file}
2. {commit subject} — {severity} — {affected file}
3. {commit subject} — {severity} — {affected file}
Prompt user for next action:
--since or --scope| Context | Command |
|---|---|
| Fix commits only | git log --format="%H %s" | grep -E "^[a-f0-9]+ fix" |
| Check test in commit | git diff-tree --no-commit-id --name-only -r {SHA} | grep -E "test|spec" |
| Files changed | git diff-tree --no-commit-id --name-only -r {SHA} |
| Fast scan | Use --quick for last 50 commits |
| Scope filter | Use --scope auth to limit to specific area |
For detailed templates, severity classification matrix, test mapping rules, and error handling, see REFERENCE.md.
Install, configure and troubleshoot MCP servers. Use when adding/enabling servers, editing .mcp.json, fixing OAuth, or when a server runs stale code after an upstream fix.
FinOps snapshot — org billing, workflow stats, cache usage. Use when you want a high-level view of CI spending or workflow health before diving deeper.
GitHub Actions billing, workflow efficiency, and waste analysis at org or repo level. Use when investigating CI/CD costs, wasted runs, or optimizing triggers.