Sloppoke documentation

The slop CLI catches AI-generated noise — defensive bloat, narrative comments, untested branches, hallucinated imports — at the commit boundary, before it lands in your repo.

Tutorials — learn by doing

For new users. Single linear path. Every step verified.

• Your first slop poke — install, log in, scan your first staged diff, apply the cleanup. ~5 minutes.

How-to — recipes for real tasks

For users who already know the basics. Pick the goal, follow the recipe.

• Install the pre-commit hook
• Gate a pull request in CI
• Use the Claude Code plugin
• Score a public GitHub repo
• Submit feedback to the learning loop
• Install from source
• Verify the binary — supply-chain checks (SHA-256, SHA256SUMS, Sigstore attestation)

Reference — facts

For users with the docs page open while working. Authoritative, predictable.

• CLI commands and flags
• Configuration (env vars + files)
• Exit codes
• Detection categories

Explanation — understanding

For users who want to know why, not just how.

• What "AI slop" actually means
• Why Sloppoke exists: LLMs are lossy compression
• How sloppoke compares — CodeRabbit, OSS slop detectors, linters
• Why pre-commit is the right boundary
• How detection works under the hood
• Privacy + identity model

Pointers outside these docs

• Product page: sloppoke.me
• Enterprise ROI model: sloppoke.me/enterprise
• Source + issue tracker: https://github.com/peeramid-labs/sloppoke
• News + release notes inside the CLI: slop news