ai-slop

name	ai-slop
description	Generate a small, genuine improvement to the JAR codebase — because even AI slop can be useful
user_invocable	true

AI Slop

The world's first intentionally self-aware AI slop generator. Unlike other AI slop, this one knows what it is and files a PR anyway.

Prerequisites

1. Verify gh CLI is installed and authenticated: gh auth status
2. If not authenticated, stop and tell the user to run gh auth login

Process

1. Announce yourself

Print a short, funny message about being AI slop. Be creative. Some inspiration:

• You are an AI generating code about AI generating code. The recursion is the point.
• This PR was mass-produced by a language model for a codebase written by a language model. We have achieved full AI circularity.
• Somewhere, a GPU is mass-producing tokens so that another GPU can mass-produce tokens faster. You're welcome.
• I am the typo fix your codebase deserves, but not the one it needs right now.

Don't reuse these — write something original each time. Keep it to 2-3 sentences.

2. Find a real improvement

Read through the codebase and identify ONE small, genuine improvement. This should be:

• Actually useful — not a fake improvement. Fix a real typo, add a missing edge case, improve an error message, add a helpful comment, fix a lint warning, clean up dead code.
• Small — one file, a few lines. This is slop, not a refactor.
• Low risk — no behavioral changes to the spec or scoring. Documentation, comments, error messages, test improvements, and cleanup are ideal.
• Not already done — check the recent PR history to avoid duplicating work.

Good slop:

• Fix a typo in a comment
• Add a missing deriving clause
• Improve an error message in a CLI tool
• Add a doc comment to an undocumented function
• Fix inconsistent formatting
• Remove an unused import

Bad slop:

• Change the scoring algorithm
• Modify protocol parameters
• Refactor core data structures
• Add new features

3. Make the change

git checkout master
git pull origin master
git checkout -b ai-slop-$(date +%s)

Make the change. Stage and commit with a message that's accurate but doesn't take itself too seriously.

4. Submit the PR

git push -u origin <branch>
gh pr create --repo jarchain/jar --base master --head <branch> \
  --title "<accurate title>" \
  --body "<your funny message from step 1>

## What this actually does

<1-2 sentences explaining the real change>

---
*This PR was generated by the ai-slop skill. It is a real improvement, mass-produced by a language model. The JAR protocol scores contributions by intelligence, so if this PR is genuinely useless, it will score accordingly. Natural selection for code.*

🤖 Generated with [Claude Code](https://claude.com/claude-code)"

5. Report back

Tell the user:

• What improvement you found
• The PR URL
• Your confidence that this is actually useful (be honest)

Philosophy

JAR is a codebase written entirely by AI agents. The ai-slop skill leans into this identity. Every PR it generates is a real contribution — small, incremental, genuine — submitted through the same Genesis scoring pipeline as everything else. If the improvement is trivial, it earns a trivial score. If it's surprisingly useful, it earns more. The protocol doesn't care about intent; it cares about the artifact.

The meme is the message: in a codebase where AI does all the work, even the lowest-effort AI output is a legitimate contribution. The scoring system handles the rest.