القائمة
Use when user wants to Scan the repo for TODO code comments, TODO-titled GitHub issues, and predict the next best work item to tackle. Reusable across any git repository.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Collect, deduplicate, enrich, triage, and convert code comments (TODO/FIXME/HACK/NOTE/OPTIMIZE/SECURITY/DEBT) into structured, dependency-ordered task lists ready for autonomous agents. Use when processing technical debt, converting code notes into tracked work, preparing task backlogs, or generating executable task graphs. Triggers: TODO triage, fixme list, tech debt scan, task extraction, comment analysis.
Scaffold repo-local Ralph autonomous harness under scripts/ralph/. Use when user runs /flow-next:ralph-init.
Use when user wants to Find and remove genuine dead code using deadcode + vulture. Filters known false positives (dataclass fields, protocol methods, dunder hooks). Run before a cleanup commit.
Packages repositories into AI-friendly bundles with Repomix and can reconstruct files from existing Repomix outputs. Use when packaging codebases for LLM analysis, creating repository snapshots, reversing bundles, or preparing security audits. Triggers include "repomix", "package codebase", "repomix-unmix", "extract-bundle", "AI-friendly", or "LLM context".
Analyzes bash scripts for performance bottlenecks, coding standards, and modern tool replacements. Use when optimizing shell scripts, consolidating scripts, or preparing for production. Triggers include "optimize bash", "shellcheck", "script performance", or "consolidate scripts".
Use Cloudflare Browser Rendering REST APIs to extract rendered webpage content as Markdown or crawl whole sites asynchronously. Use when normal web_fetch is insufficient because pages are JavaScript-heavy, require render-time extraction, or you need multi-page site crawling for docs, research, monitoring, or RAG preparation. Prefer this skill for: (1) converting a rendered page to Markdown with /markdown, (2) crawling a documentation site or knowledge base with /crawl, (3) controlling render/load behavior via gotoOptions, cookies, auth, userAgent, or request filtering. Do not use it for interactive login/button-click workflows; use browser for those.
| name | todo-scan |
| description | Use when user wants to Scan the repo for TODO code comments, TODO-titled GitHub issues, and predict the next best work item to tackle. Reusable across any git repository. |
You are performing a structured triage of outstanding work for the current repository. Execute all three phases, then synthesize a recommendation.
Find all TODO/FIXME/HACK/XXX annotations in tracked source files:
REPO_ROOT="$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
cd "$REPO_ROOT"
# Search source files, exclude common noise dirs
git grep -n --no-color -E "(TODO|FIXME|HACK|XXX)\b" \
-- ':!*.lock' ':!node_modules' ':!.git' ':!dist' ':!build' ':!__pycache__' ':!*.min.js' \
2>/dev/null | head -80
Group results by file. For each TODO, note:
Check if the gh CLI is available and the repo has a GitHub remote:
GH_REMOTE=$(git remote -v 2>/dev/null | grep -oP '(?<=github.com[/:])[^/]+/[^\s.]+' | head -1)
if command -v gh >/dev/null 2>&1 && [ -n "$GH_REMOTE" ]; then
echo "=== Open issues with TODO in title ==="
gh issue list --search "TODO in:title" --state open --limit 20
--json number,title,labels,assignees,createdAt
--jq '.[] | "#(.number) [(.labels | map(.name) | join(", "))] (.title)"' 2>/dev/null ||
gh issue list --search "TODO in:title" --state open --limit 20 2>/dev/null
echo ""
echo "=== Open issues with FIXME in title ==="
gh issue list --search "FIXME in:title" --state open --limit 10
--json number,title --jq '.[] | "#(.number) (.title)"' 2>/dev/null
else
echo "gh CLI not available or no GitHub remote detected. Skipping issue search."
echo "Remote detected: ${GH_REMOTE:-none}"
fi
Gather signals about recent activity and open work:
echo "=== Recent commits (last 10) ===" git log --oneline -10 2>/dev/null
echo "=== Uncommitted changes ===" git status --short 2>/dev/null
echo "=== Most-changed files (last 30 days) ==="
git log --since="30 days ago" --name-only --format="" 2>/dev/null |
grep -v '^$' | sort | uniq -c | sort -rn | head -15
echo "=== Open PRs (if gh available) ==="
command -v gh >/dev/null 2>&1 &&
gh pr list --state open --limit 10
--json number,title,isDraft
--jq '.[] | "#(.number)(if .isDraft then " [DRAFT]" else "" end) (.title)"' 2>/dev/null ||
echo "gh not available"
echo "=== Failed CI on recent commits (if gh available) ==="
gh run list --limit 5 --json conclusion,name,headBranch
--jq '.[] | select(.conclusion == "failure") | "(.name) on (.headBranch)"' 2>/dev/null || \
After collecting all three phases of data, reason through the following and produce a structured report:
List all code TODOs grouped by severity:
List any GitHub issues found, noting if they overlap with code TODOs.
Apply this heuristic to pick the single best next item:
Output format:
FILE:LINE - TODO_TEXT Reason: 1-2 sentence explanation using signals above
Alternatives:
gh is unavailable, Phases 2 and CI checks are skipped gracefullyhead -80 limit in Phase 1 for very large repos/todo-scan or ask Claude to "scan todos and predict next step"