Habitusx — AI skill for Claude Code
The AI Code Outcomes Index.
How to install Habitusx
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ptengelmann/habitusx and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Habitusx does
The AI Code Outcomes Index. What happens to AI-written code after it lands: revert rate, PR acceptance and survival by agent, language and repo, against a human baseline. Neutral, continuous, methodology published.
Alternatives in AI
- Vllm Mlx — High-performance OpenAI and Anthropic compatible LLM inference server for Apple Silicon 1.5k ★
- Devo — Model-neutral agent desktop/runtime for private, enterprise, and OpenAI-compatible / Anthropic-compatible mode 316 ★
- SoloDawn — Catches and repairs AI hallucinations before delivery: three quality gates, 31 built-in rules, a self-healing 312 ★
README
HabitusX
**The AI Code Outcomes Index.** What happens to AI-written code after it lands.
Millions of public GitHub commits now carry markers from AI coding agents. Existing trackers count them. HabitusX measures what happens next: revert rate, pull request acceptance, review burden and survival, by agent, language and repository, against a human baseline from the same repositories. Continuous, neutral, methodology published.
Repository layout
backend/ Python pipeline, attribution engine, API see backend/README.md
frontend/ Next.js site (not started; UI phase)
registry/ community-contributable agent detection rules see registry/README.md
docs/ architecture, methodology, roadmap, ADRs
Status
PR 1 (foundation) delivers the deterministic core: git trailer parsing, revert detection, a data-driven attribution engine, a validated registry of ten agents, and the full quality gate in CI. Data ingestion from GitHub Archive is PR 2. See `docs/roadmap.md`.
Develop
cd backend
uv sync
uv run python -m habitusx.cli --help
uv run python -m habitusx.cli registry validate
uv run ruff format --check . ; uv run ruff check . ; uv run mypy ; uv run pytest --cov
Try the engine on a message:
"Fix retry`n`nCo-Authored-By: Claude " | uv run python -m habitusx.cli attribute
Contributing
See [CONTRIBUTING.md](CONTRIBUTING.md). Every change goes through a pull request with CI green. Commit messages follow Conventional Commits (`feat:`, `fix:`, `docs:`, ...). Architectural decisions get an ADR in `docs/adr/`. Registry contributions: `registry/README.md`.
License
Code: [Business Source License 1.1](LICENSE), converting to Apache 2.0 on 2029-09-05. You may run it, modify it, and analyse your own repositories with it; you may not offer it as a competing hosted service before the change date. Registry data: [CC BY 4.0](registry/LICENSE).
Principles
Pure, tested core. Explaina
Related Skills
SnapAI
Fast macOS screenshot & screen-recording tool written in Rust, built for handing captures to AI agents (Claude
Sec Scan Auth
인증/인가/어뷰징 취약점 진단 — Auth Bypass / IDOR / Mass Assignment / Rate Limit / 멱등성 / 클라이언트 신뢰 로직 — scan_api.py + scan_
Team Foundry
One shared context layer for Claude Code, Cursor, Gemini CLI, and Codex — so the whole team's AI tools build f
Cc Judge
Shared LLM-evals framework for running Claude Code agents in containers, scoring outcomes with an LLM judge, a
Chunsun
Self-deployed AI-native project management platform—requirements, acceptance, and progress stay in your instan
Canopus
Canopus keeps long-running AI coding agents locked on their goal: frozen acceptance envelopes, deterministic c
Related Agents
Selfish Audience Intelligence
Build an evidence-based model of the people the user can help, the problems they experience, and the language/
Voc Miner
Extracts and categorises verbatim voice-of-customer language (pains, desired outcomes, triggers, objections, c
Iteration Guard
The verify gate of the Logic-Lens iteration loop. Given a baseline iteration and a candidate iteration, compar