The open-source CLI that scores AI-generated code in under a second, and the PR gate that keeps slop out of main. Every finding goes back to the agent that wrote it.
See what is passing, what needs attention, and where to act before AI-written code reaches production.
4 repos, 3 passing, 1 failing, 75 avg
The same deterministic scan runs on your laptop, on every pull request, and inside your agent loop. Same code in, same score out.
One command, no account, nothing leaves your machine. 50+ rules across 11 language targets.
$ npx aislop@latest scanscanaislop scores every pull request as a GitHub check and holds merges below your threshold.
aislop prepares the fix prompt for Claude Code, Cursor, Codex, and 11 more. Your reviewers get a clean diff.
+ any agent via --prompt
Goose
Kimi CLIFree for public repositories after an instant scan. Contributors and reviewers see the live quality score before they trust a generated PR.
Your quality bar, at a glance
https://badges.scanaislop.com/score/<owner>/<repo>.svg[](https://scanaislop.com/<owner>/<repo>)
Every rule comes from a measured failure mode in real repositories, benchmark papers, or regression tests.
We scanned popular TypeScript, Python, Go, Rust, Ruby, PHP, and Java projects, then cut noisy findings by 38% without disabling rules.
Read reportSlopCodeBench measured verbosity and structural erosion. We translated the repeatable Python signals into four shipped rules.
Read reportThe first public scan run exposed the shapes that still drive the roadmap: narrative comments, dead code, unsafe casts.
Read reportAnother slop code example, run through a real scan: the agent-written TypeScript, the 14 findings aislop 0.20.0 reported, what aislop fix removed, and what still needed a person.
Read moreAn AI slop website example taken apart: the generic hero, gradient blob and placeholder testimonial on the page, the dead CSS and swallowed errors in the code, and the fixes.
Read moreRun bunx aislop scan or npx aislop scan, read the score, then use aislop scan --share (new in 0.20.0) to get a public report link and README badge for a GitHub repo.
Read moreaislop is a code quality CLI for AI generated code. Run npx aislop scan to score a repo 0 to 100 for AI slop: swallowed exceptions, narrative comments, unsafe casts, dead residue. Same code in, same score out.
AI slop is low-value residue in AI generated code. It compiles and often passes tests, then still costs review. Also searched as slop code or code slop.
Slop code is AI slop in a repository: empty catches, leftover TODOs, unused helpers, comments that restate the next line. aislop names those patterns and can gate them in CI.
Run tests and a deterministic scan first. npx aislop scan is the aislop path. Then review intent, not formatting.
Start with one CLI command. Choose a plan when your team is ready to gate pull requests, track drift, and hand fixes back to agents.