project-graveyard

Autopsy abandoned git repositories to diagnose why side projects died.

135k|19.9k|Updated Apr 29, 2024
One-click install
npx skills add https://github.com/Shubhamsaboo/awesome-llm-apps --skill project-graveyard
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: project-graveyard
Source: https://github.com/Shubhamsaboo/awesome-llm-apps/tree/main/agent_skills/project-graveyard
Command: npx skills add https://github.com/Shubhamsaboo/awesome-llm-apps --skill project-graveyard

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Developers accumulate folders of unfinished side projects without understanding why each one stopped. This Skill reads local git history to diagnose the cause of death for every abandoned repo, surface personal abandonment patterns, and identify which corpse is closest to shipping.

Core Features & Use Cases

  • Cause-of-Death Autopsy: Classifies each dead repo into categories like shiny_object, deploy_fear, payments_wall, auth_wall, boilerplate_wall, rewrite_spiral, scope_explosion, or slow_fade based on forensic git evidence.
  • Pattern Detection: Identifies recurring developer habits such as projects dying at a specific day count or being killed by newer side projects.
  • Pulse Ranking: Scores each corpse 0-100 on resurrectability based on README presence, tests, commit momentum, and deploy readiness.
  • Resurrection Planning: Generates a 7-step plan ending at shipped, with a world-check for APIs or tools that may have made the original blockers easier.
  • Relapse Watch: Tracks previously resurrected projects across scans and reports whether they are holding or going silent again.
  • Use Case: A developer asks "why do I never finish my side projects?" The agent scans ~/dev and ~/projects, produces a tombstone report naming each corpse with its cause, surfaces the pattern that four of six projects died at day 19, and offers to resurrect the strongest pulse candidate.

Quick Start

Run the graveyard on my ~/dev and ~/projects folders and tell me which project is closest to shipping.

Frequently Asked Questions about project-graveyard

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find abandoned side projects on my machine?

Run python3 scripts/graveyard.py ~/dev ~/projects to scan those directories for git repositories silent for more than 45 days. The script walks up to 4 levels deep, skips node_modules and .venv, and reports each dead repo with its cause of death and pulse score.

How to detect why a git project was abandoned?

The graveyard.py script reads commit messages, file touches, and timing from git log. It classifies deaths into categories like shiny_object (killed by a newer repo), deploy_fear (finished but never shipped), or payments_wall based on forensic evidence from the history.

Does the graveyard scanner require network access?

No. The script uses only Python stdlib modules and runs entirely offline. It reads git metadata locally and never makes network calls, so scanned code contents never leave the machine.

Why are some of my repos skipped as not-yours?

The scanner filters by git commit email. Repos where you authored less than 20% of commits are treated as clones or work checkouts. Pass --me [email protected] to claim commits under additional identities, or --include-foreign to scan everything.

What does the pulse score measure for resurrection?

Pulse is a 0-100 resurrectability score based on README presence, test coverage, config-to-code ratio, commit count, remote existence, recency, and deploy config. Higher scores indicate corpses closer to shipping with less remaining work.

When should I not use the project graveyard skill?

Do not use it for cleaning disk space (use kondo or npkill), archiving GitHub repos, or deep analysis of a single repo's history. The skill is designed for surveying many abandoned projects at once and identifying resurrection candidates.