auto-enforcer-actionlisted
Install: claude install-skill Habitat-Thinking/ai-literacy-superpowers
# Auto-Enforcer GitHub Action
## Overview
The harness enforces constraints across three loops:
- **Inner loop** (PreToolUse hooks) — advisory, runs while editing in
Claude Code
- **Middle loop** (CI on PR) — runs automatically on every pull request
- **Outer loop** (scheduled audit) — investigative, runs on demand via
`/harness-audit`
The standard `harness.yml` CI workflow handles the middle loop for
deterministic constraints only. Agent-based PR constraints — rules like
"All frontmatter has name and description" that require judgement —
are silently skipped at PR time unless someone remembers to run
`/harness-audit` manually.
The auto-enforcer closes this gap. It runs every PR-scoped constraint
in HARNESS.md automatically on each pull request:
- **Deterministic constraints** execute their tool command directly.
Failures are blocking — they fail the CI job and prevent merge.
- **Agent constraints** send the constraint rule plus the PR diff to
the Claude API. Findings are advisory — posted as a PR comment but
never blocking. AI judgement should inform reviewers, not gate merges.
The workflow is data-driven: it reads constraint definitions from
HARNESS.md at runtime. Adding or modifying constraints in HARNESS.md
changes what the action checks with no workflow edits needed.
---
## Prerequisites
1. **HARNESS.md** exists at the project root with at least one
PR-scoped constraint
2. **ANTHROPIC_API_KEY** stored as a GitHub Actions secret:
Settings > Secret