Skip to main content

ag-openclaw-github-repo-commander

7-stage super workflow for GitHub repo audit, cleanup, PR review, and competitor analysis

Informations de source

Dépôt
GrayCodeAI/starling
Dernière activité de la source
13 juillet 2026 à 09:28
Langue détectée de SKILL.md
anglais
Étoiles
2
Forks
0

Options d'installation

Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.

Vérifiez les fichiers source

Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.

Affichage de SKILL.md

SKILL.md
Instructions source · Aperçu en lecture seule
name
ag-openclaw-github-repo-commander
description
7-stage super workflow for GitHub repo audit, cleanup, PR review, and competitor analysis
license
MIT
tags
["github","git","repository","audit","cleanup"]
category
development-and-testing
risk
safe
source
community
date_added
2026-03-18T00:00:00.000Z
author
wd041216-bit
tools
["claude","cursor"]
# OpenClaw GitHub Repo Commander ## Overview A structured 7-stage super workflow for comprehensive GitHub repository management. This skill automates repository auditing, cleanup, competitor benchmarking, and optimization — turning a messy repo into a clean, well-documented, production-ready project. ## When to Use This Skill - Use when you need to audit a repository for secrets, junk files, or low-quality content - Use when the user says "clean up my repo", "optimize my GitHub project", or "audit this library" - Use when reviewing or creating pull requests with structured analysis - Use when comparing your project against competitors on GitHub - Use when running `/super-workflow` or `/openclaw-github-repo-commander` on a repo URL ## How It Works ### Stage 1: Intake Clone the target repository, define success criteria, and establish baseline metrics. ### Stage 2: Execution Run `scripts/repo-audit.sh` — automated checks for: - Hardcoded secrets (`ghp_`, `sk-`, `AKIA`, etc.) - Tracked `node_modules/` or build artifacts - Empty directories - Large files (>1MB) - Missing `.gitignore` coverage - Broken internal README links ### Stage 3: Reflection Deep manual review beyond automation: content quality, documentation consistency, structural issues, version mismatches. ### Stage 4: Competitor Analysis Search GitHub for similar repositories. Compare documentation standards, feature coverage, star counts, and community adoption. ### Stage 5: Synthesis Consolidate all findings into a prioritized action plan (P0 critical / P1 important / P2 nice-to-have). ### Stage 6: Iteration Execute the plan: delete low-value files, fix security issues, upgrade documentation, add CI workflows, update changelogs. ### Stage 7: Validation Re-run the audit script (target: 7/7 PASS), verify all changes, push to GitHub, and deliver a full report. ## Examples ### Example 1: Full Repo Audit ``` /openclaw-github-repo-commander https://github.com/owner/my-repo ``` Runs all 7 stages and produces a detailed before/after report. ### Example 2: Quick Cleanup ``` Clean up my GitHub repo — remove junk files, fix secrets, add .gitignore ``` ### Example 3: Competitor Benchmarking ``` Compare my skill repo with the top 5 similar repos on GitHub ``` ## Best Practices - ✅ Always run Stage 7 validation before pushing - ✅ Use semantic commit messages: `chore:`, `fix:`, `docs:` - ✅ Check the `pr_todo.json` file for pending reviewer requests - ❌ Don't skip Stage 4 — competitor analysis reveals blind spots - ❌ Don't commit `node_modules/` or `.env` files ## Security & Safety Notes - The audit script scans for common secret patterns but excludes `.github/workflows/` to avoid false positives - All `gh` CLI operations use the user's existing authentication — no credentials are stored by this skill - The skill never modifies files without explicit user confirmation in Stage 6 ## Source Repository [github.com/wd041216-bit/openclaw-github-repo-commander](https://github.com/wd041216-bit/openclaw-github-repo-commander) **License**: MIT | **Version**: 4.0.0 ## Limitations - Use this skill only when the task clearly matches the scope described above. - Do not treat the output as a substitute for environment-specific validation, testing, or expert review. - Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.
Voir sur GitHub