slopmop 0.4.1

Quality gates for AI-assisted codebases — catch the slop LLMs leave behind.

🪣 Slop-Mop

Quality gates for AI-assisted codebases. Not a silver bullet — just a mop.

The code generated by frontier LLMs shares the same features. These days, the majority of those features are positive, but there are still remnants from agent training that result in a set of undesirable outcomes from agentic LLMs — especially when working with very small development teams or "vibe coded" projects.

The frontier agents used to generate code were all effectively trained to "accomplish the task at any cost." This makes them great at accomplishing small things with blinders on, but dangerous for the long-term stability of a repository.

The modern development process has turned into a game of whack-a-mole where we're trying to use the output of the agents while also catching LLMs in instances of shortsightedness, deception, overconfidence, and laziness. Instead of repeating yourself in chat ad nauseam, let slop-mop do that steering for you.

Not only will slop-mop catch many of the most egregious examples of these behaviors, but it will also tell the agent the precise mistake made and exactly how to fix it. This is a new approach for agentic development — turning a "bug" (relentless task accomplishment) into a feature (relentless code quality improvement).

---

Quick Start

# Install (once per machine)
pipx install slopmop          # recommended — isolated, no dep conflicts
# or: pip install slopmop

# Set up your project
sm init                       # auto-detects languages, writes .sb_config.json

# Run quality gates
sm validate commit            # fix what it finds, commit when green

sm init auto-detects Python, JavaScript, or both and writes a .sb_config.json with applicable gates enabled.

---

The Loop

Development with slop-mop follows a single repeated cycle:

sm validate commit → see what fails → fix it → repeat → commit

When a gate fails, the output tells you exactly what to do next:

┌──────────────────────────────────────────────────────────┐
│ 🤖 AI AGENT ITERATION GUIDANCE                           │
├──────────────────────────────────────────────────────────┤
│ Profile: commit                                          │
│ Failed Gate: deceptiveness:py-coverage                   │
├──────────────────────────────────────────────────────────┤
│ NEXT STEPS:                                              │
│                                                          │
│ 1. Fix the issue described above                         │
│ 2. Validate: sm validate deceptiveness:py-coverage       │
│ 3. Resume:   sm validate commit                          │
│                                                          │
│ Keep iterating until all the slop is mopped.             │
└──────────────────────────────────────────────────────────┘

This is purpose-built for AI agents. The guidance is machine-readable, the iteration is mechanical, and the agent never has to wonder what to do next. It saves tokens (no flailing), saves CI dollars (catch it locally), and keeps the codebase habitable long-term.

Use sm status for a report card of all gates at once.

---

Why These Categories?

Gates aren't organized by language — they're organized by the failure mode they catch. These are the four ways LLMs reliably degrade a codebase:

🔴 Overconfidence

"It compiles, therefore it's correct."

The LLM generates code that looks right, passes a syntax check, and silently breaks at runtime. These gates verify that the code actually works.

Gate	What It Does
overconfidence:py-tests	🧪 Runs pytest — code must actually pass its tests
overconfidence:py-static-analysis	🔍 mypy strict — types must check out
overconfidence:py-types	🔬 pyright strict — second opinion on types
overconfidence:js-tests	🧪 Jest test execution
overconfidence:js-types	🏗️ TypeScript type checking (tsc)
overconfidence:deploy-script-tests	🚀 Validates deploy scripts

🟡 Deceptiveness

"Tests pass, therefore the code is tested."

The LLM writes tests that assert nothing, mock everything, or cover the happy path and call it done. Coverage numbers look great. The code is still broken.

Gate	What It Does
deceptiveness:py-coverage	📊 Whole-repo coverage (80% default threshold)
deceptiveness:py-diff-coverage	📈 Coverage on changed lines only (diff-cover)
deceptiveness:bogus-tests	🧟 AST analysis for tests that assert nothing
deceptiveness:js-coverage	📊 JavaScript coverage analysis
deceptiveness:js-bogus-tests	🎭 Bogus test detection for JS/TS

🟠 Laziness

"It works, therefore it's done."

The LLM solves the immediate problem and moves on. Formatting is inconsistent, dead code accumulates, complexity creeps upward, and nobody notices until the codebase is incomprehensible.

Gate	What It Does
laziness:py-lint	🎨 autoflake, black, isort, flake8 (supports auto-fix 🔧)
laziness:js-lint	🎨 ESLint + Prettier (supports auto-fix 🔧)
laziness:complexity	🌀 Cyclomatic complexity (max rank C)
laziness:dead-code	💀 Dead code detection via vulture (≥80% confidence)
laziness:template-syntax	📄 Jinja2 template validation
laziness:js-frontend	⚡ Quick ESLint frontend check

🔵 Myopia

"My change is fine. Why would I look at the bigger picture?"

The LLM has a 200k-token context window and still manages tunnel vision. It duplicates code across files, ignores security implications, and lets functions grow unbounded because it can't see the pattern.

Gate	What It Does
myopia:loc-lock	📏 File and function length limits
myopia:source-duplication	📋 Code clone detection (jscpd)
myopia:string-duplication	🔤 Duplicate string literal detection
myopia:security-scan	🔐 bandit + semgrep + detect-secrets
myopia:security-audit	🔒 Full security audit (code + pip-audit)

PR Gates

Gate	What It Does
pr:comments	💬 Checks for unresolved PR review threads

---

Profiles

Profiles bundle gates into workflows. Use profiles, not individual gates:

Profile	Gates	When to Use
commit	17 gates — all overconfidence, deceptiveness, laziness, myopia checks	Before every commit
pr	19 gates — all commit gates + PR comments + diff-coverage	Before opening or updating a PR
quick	2 gates — lint + security scan	Fast feedback during development
python	5 gates — Python-specific subset	Language-focused validation
javascript	5 gates — JS/TS-specific subset	Language-focused validation
quality	5 gates — complexity, duplication, loc-lock	Code quality only
security	1 gate — full security audit	Security-focused validation

JS gates auto-skip when no JavaScript is detected.

---

Getting Started: The Remediation Path

Most projects won't pass all gates on day one. That's expected. Here's the ramp:

1. Initialize

sm init                       # auto-detects everything, writes .sb_config.json

2. See Where You Stand

sm validate commit            # run all gates, see what fails
sm status                     # full report card

3. Disable What You're Not Ready For

sm config --disable laziness:complexity        # too many complex functions right now
sm config --disable deceptiveness:py-coverage  # coverage is at 30%, not 80%
sm validate commit                             # get the rest green first

4. Fix Everything That's Left

Iterate: run sm validate commit, fix a failure, run again. The iteration guidance tells you exactly what to do after each failure.

5. Install Hooks

sm commit-hooks install commit    # pre-commit hook runs quality gates
sm commit-hooks status            # verify hooks are installed

Now every git commit runs slop-mop. Failed gates block the commit.

6. Re-enable Gates Over Time

sm config --enable laziness:complexity         # refactored enough, turn it on
sm config --enable deceptiveness:py-coverage   # coverage is at 75%, set threshold to 70

7. Let Agents Vibe-Code

With hooks in place, agents can write code freely. Slop-mop catches the slop before it reaches the repo. This saves tokens (no back-and-forth debugging), saves CI money (catch it locally), and keeps the codebase survivable long-term.

---

Configuration

sm config --show              # show all gates and their status
sm config --enable <gate>     # enable a disabled gate
sm config --disable <gate>    # disable a gate
sm config --json <file>       # bulk update from JSON

Include / Exclude Directories

sm config --exclude-dir myopia:generated       # skip generated code
sm config --include-dir overconfidence:src      # only check src/

• include_dirs: whitelist — only these dirs are scanned
• exclude_dirs: blacklist — always skipped, takes precedence

.sb_config.json

Edit directly for per-gate configuration:

{
  "version": "1.0",
  "default_profile": "commit",
  "python": {
    "gates": {
      "coverage": { "threshold": 80 },
      "tests": { "test_dirs": ["tests"] }
    }
  },
  "quality": {
    "exclude_dirs": ["generated", "vendor"]
  }
}

---

CI Integration

GitHub Actions

name: slop-mop
on:
  pull_request:
    branches: [main]
  push:
    branches: [main]

jobs:
  validate:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
      - uses: actions/setup-python@v5
        with:
          python-version: '3.11'
      - run: pip install slopmop
      - run: sm validate commit
      - if: github.event_name == 'pull_request'
        env:
          GH_TOKEN: ${{ github.token }}
        run: sm validate pr:comments

Check CI Status Locally

sm ci               # current PR
sm ci 42             # specific PR
sm ci --watch        # poll until CI completes

---

Architecture

Slop-mop installs as a normal package and is configured per-project via .sb_config.json. The sm command is on your PATH once and works in any repo.

Tool resolution order — sm uses your project's tools when available:

1. <project_root>/venv/bin/<tool> or .venv/bin/<tool> — project-local venv (highest priority)
2. $VIRTUAL_ENV/bin/<tool> — currently activated venv
3. System PATH — sm's own bundled tools (via pipx)

This means if your project has its own pytest (with plugins like pytest-django), sm uses it. Otherwise, sm falls back to its own.

Submodule alternative: If you need strict version pinning, add slop-mop as a git submodule and invoke python -m slopmop.sm directly. Supported but not recommended for most projects.

---

Development

# Working on slop-mop itself
pip install -e .
sm validate --self            # dogfooding — sm validates its own code
pytest

See CONTRIBUTING.md for the process of adding new gates.

---

Further Reading

📖 A Hand for Daenerys: Why Tyrion Is Missing from Your Vibe-Coding Council — the article that started this project.

---

License

Slop-Mop Attribution License v1.0 — free to use, modify, and redistribute with attribution.

P.S. Other than this line in the readme and a few scattered lines here and there, nothing in this project was written by a human. It is, for better or worse, the result of living under the slop-mop regime.