slopcheck 0.3.0

Detect AI-hallucinated packages before you install them.

slopcheck

Detect AI-hallucinated packages before you install them.

When your AI coding assistant suggests flask-gpt-helper or easy-requests, those packages probably don't exist. But someone might register them as malware before you notice. That's slopsquatting.

slopcheck catches it first.

Install

pip install slopcheck

Or one-liner if you're in a hurry:

Mac/Linux:

curl -fsSL https://raw.githubusercontent.com/0xToxSec/slopcheck/main/install.sh | bash

Windows (PowerShell):

irm https://raw.githubusercontent.com/0xToxSec/slopcheck/main/install.ps1 | iex

Usage

Scan your project

# Auto-detect dependency files in current directory
slopcheck .

# Scan a specific file
slopcheck requirements.txt

Safe install (check first, install if clean)

# Instead of: pip install flask requests sketchy-package
slopcheck install flask requests sketchy-package

# Auto-detects ecosystem from your project (package.json = npm, etc.)
# Or force it:
slopcheck install express lodash --ecosystem npm

# Install suspicious packages anyway (slop is ALWAYS blocked):
slopcheck install some-package --force

Slop gets blocked. Always. Suspicious packages get skipped unless you pass --force. Clean packages install normally through your real package manager.

Check a single package

slopcheck flask-gpt-helper --pkg pypi
slopcheck react-ai-utils --pkg npm
slopcheck easy-http --pkg crates.io
slopcheck github.com/fake/module --pkg go

Output

  [SLOP] flask-gpt-helper (pypi)
    > Package 'flask-gpt-helper' does not exist on pypi. Your AI made it up.
    > Name ends with '-helper' -- classic LLM naming pattern

  [SLOP] reqeusts (pypi)
    > Package 'reqeusts' does not exist on pypi. Your AI made it up.
    ? Did you mean: requests

  [SUS] easy-requests (pypi)
    > Name starts with 'easy-' -- classic LLM naming pattern. Package exists but the name screams 'LLM bait'.

  [OK] requests (pypi)

JSON output (for CI)

slopcheck requirements.txt --json

What it detects

• Non-existent packages -- the #1 signal. If it's not on the registry, your AI made it up.
• Brand new packages -- created in the last 7 days? Probably registered to trap you.
• Low downloads -- under 100 downloads means nobody uses it.
• Hallucination patterns -- LLMs love naming packages {popular-lib}-{ai|gpt|helper|utils}. We check for these patterns.
• Typosquats -- Levenshtein distance check against popular packages with "did you mean?" suggestions.
• Missing repo links -- legitimate packages almost always link to source code.

Supported ecosystems

Ecosystem	Dependency files	Registry
PyPI	requirements.txt, pyproject.toml, Pipfile, Pipfile.lock	pypi.org
npm	package.json	npmjs.org
crates.io	Cargo.toml	crates.io
Go	go.mod	proxy.golang.org

Exit codes

Code	Meaning
0	Clean -- all packages check out
1	Suspicious -- some packages deserve a second look
2	Slop detected -- hallucinated or dangerously new packages found

Options

slopcheck [target] [options]

target          Directory, file, or package name (default: .)
--pkg ECOSYSTEM Check single package (pypi, npm, crates.io, go)
--workers N     Parallel registry checks (default: 10)
--json          JSON output for CI pipelines

GitHub Action

Add this to your repo at .github/workflows/slopcheck.yml and every PR that touches dependency files gets scanned automatically:

name: slopcheck

on:
  pull_request:
    paths:
      - 'requirements*.txt'
      - 'pyproject.toml'
      - 'package.json'
      - 'Cargo.toml'
      - 'go.mod'

jobs:
  slopcheck:
    runs-on: ubuntu-latest
    permissions:
      pull-requests: write
    steps:
      - uses: actions/checkout@v4
      - uses: 0xToxSec/slopcheck@main
        with:
          path: '.'
          fail-on: 'slop'

If slop is found, the action fails the check and drops a comment on the PR with the full report. Set fail-on: 'sus' to be stricter.

License

MIT