G
gretchenfitze/anti-slop
Agent skills that keep AI-written TypeScript and React code in check: measure which anti-slop checks a repo…
no SECURITY.md
View on GitHub ↗Link checked 2026-11-12 · HTTP 200 · opens the vendor's own site
Install & config
Client config (built from the install command this repo gives)
{
"mcpServers": {
"gretchenfitze-anti-slop": {
"command": "npx",
"args": ["skills"]
}
}
}Parsed from the README we fetched on 2026-10-26; we have not executed it.
command/args come from the install command the repository states. When it is ambiguous, the project README wins.
VS Code (.vscode/mcp.json - the key is servers)
{
"servers": {
"gretchenfitze-anti-slop": {
"command": "npx",
"args": ["skills"]
}
}
}Same server, but VS Code keeps it under <code>servers</code>.
Claude Code (command line)
claude mcp add --transport stdio gretchenfitze-anti-slop -- npx skills
Claude Code does not read a config file; add the server with this command.
Key names differ by client: Clients & CLI
- Star count
- 4 ★ · gretchenfitze/anti-slop · read from the GitHub search API · (checked 2026-11-13 20:00:50)
- Source
- read directly from its GitHub page · type: open-source project
- Harvest history
- first seen 2026-10-21 · last seen 2026-11-14 · seen 28× · listed by 1 sources
- Score
- score 103 · 4 ★ · 1 sources · repo
- Install
npx skills
FAQ (answers come from our records)
How do I install it?
The repository provides this install command:
npx skills. It comes from the repository itself; we cannot guarantee it runs as-is in your client.Who is it for?
We filed it under these tags: Agent 技能. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
gretchenfitze/anti-slop；first catalogued 2026-10-21；last seen 2026-11-14。There is no official certification or quality rating here — we record sources and dates only.# Anti-slop
以下摘录自该仓库的 README（片段） · 在 GitHub 读完整 README →
Two agent skills that harden a TypeScript or React repo against AI-generated slop. One measures which checks a project actually needs and plans them as small PRs; the other rolls the plan out one PR at a time.
Companion to 10 anti-AI slop moves for frontend projects going faster than humans can review by Yuri Mikhin.
These skills are built by Evil Martians , an American design and engineering consultancy for developer tools, AI, and cybersecurity startups .
## The skills
| Skill | What it does | Changes |
| anti-slop-audit | Runs every check read-only (tsc flags, oxlint with sonarjs and effect rules, knip, jscpd, react-doctor), removes framework false positives, mines past bugs for custom lint rules, and writes a PR-by-PR plan | anti-slop/plan.md and anti-slop/todo.md only |
| anti-slop-rollout | Takes the next step of the plan: adds one check, fixes or baselines what it finds, documents the commands for agents and humans, verifies, opens the PR | one branch and PR per run |
The audit judges each move against the project instead of applying all ten. It skips the API contract on a site without an API, defers mutation testing when the test runner has no Stryker runner, and separates the dead files knip reports from the ones that are actually dead.
It covers three stages:
An existing codebase gets a gradual rollout: warnings on changed files first, CI gates last.
A prototype turning into a product gets a plan scoped to the code that survives, with new slop stopped before the old gets cleaned.
A project just starting gets a short day-one plan with every check as an error, since there's nothing to clean up yet. If the stack is Vite + React or NestJS, Yuri Mikhin's ts-workbench template is often the faster start, and the audit says so.
The audit reads the stack from the repo (linter, test runner, e2e, API, package manager, git hooks, agents, CI) and picks tools to match: rule sets for ESLint, oxlint, or Biome, Jest or Vitest mock checks, the project's package manager in every command.
Four choices have two reasonable answers. The plan takes option 1, what the article recommends, and the audit's report asks whether to switch any:
| Choice | Option 1 (the article) | Option 2 |
| Lint backlog | Gradual: warnings on changed files, old code left as-is, a rule promoted once it catches a real bug | Cleanup: PRs that fix the backlog to zero, no warn left |
| Mutation testing in CI | Required, with no score threshold | Report-only, never blocks a PR |
| exactOptionalPropertyTypes | Always on | Count-gated: on only when the error count is small |
| react-doctor (React) | Report: a command and a scheduled run | Gated: errors cleaned up, then blocking |
## Install
Install both skills: the rollout runs the audit's probe from the folder next to it.
sh
npx skills add gretchenfitze/anti-slop --skill '' -a claude-code -g
Pass your agent to -a: claude-code, codex, cursor, gemini-cli, or github-copilot. Drop -g to install into the current project instead of your home directory.
Claude Code can install the bundle from this repo's plugin marketplace instead:
Companion to 10 anti-AI slop moves for frontend projects going faster than humans can review by Yuri Mikhin.
These skills are built by Evil Martians , an American design and engineering consultancy for developer tools, AI, and cybersecurity startups .
## The skills
| Skill | What it does | Changes |
| anti-slop-audit | Runs every check read-only (tsc flags, oxlint with sonarjs and effect rules, knip, jscpd, react-doctor), removes framework false positives, mines past bugs for custom lint rules, and writes a PR-by-PR plan | anti-slop/plan.md and anti-slop/todo.md only |
| anti-slop-rollout | Takes the next step of the plan: adds one check, fixes or baselines what it finds, documents the commands for agents and humans, verifies, opens the PR | one branch and PR per run |
The audit judges each move against the project instead of applying all ten. It skips the API contract on a site without an API, defers mutation testing when the test runner has no Stryker runner, and separates the dead files knip reports from the ones that are actually dead.
It covers three stages:
An existing codebase gets a gradual rollout: warnings on changed files first, CI gates last.
A prototype turning into a product gets a plan scoped to the code that survives, with new slop stopped before the old gets cleaned.
A project just starting gets a short day-one plan with every check as an error, since there's nothing to clean up yet. If the stack is Vite + React or NestJS, Yuri Mikhin's ts-workbench template is often the faster start, and the audit says so.
The audit reads the stack from the repo (linter, test runner, e2e, API, package manager, git hooks, agents, CI) and picks tools to match: rule sets for ESLint, oxlint, or Biome, Jest or Vitest mock checks, the project's package manager in every command.
Four choices have two reasonable answers. The plan takes option 1, what the article recommends, and the audit's report asks whether to switch any:
| Choice | Option 1 (the article) | Option 2 |
| Lint backlog | Gradual: warnings on changed files, old code left as-is, a rule promoted once it catches a real bug | Cleanup: PRs that fix the backlog to zero, no warn left |
| Mutation testing in CI | Required, with no score threshold | Report-only, never blocks a PR |
| exactOptionalPropertyTypes | Always on | Count-gated: on only when the error count is small |
| react-doctor (React) | Report: a command and a scheduled run | Gated: errors cleaned up, then blocking |
## Install
Install both skills: the rollout runs the audit's probe from the folder next to it.
sh
npx skills add gretchenfitze/anti-slop --skill '' -a claude-code -g
Pass your agent to -a: claude-code, codex, cursor, gemini-cli, or github-copilot. Drop -g to install into the current project instead of your home directory.
Claude Code can install the bundle from this repo's plugin marketplace instead:
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/gretchenfitze-anti-slop/)
- Source
- github · https://github.com/gretchenfitze/anti-slop
- Browse by field
- Agent skills
- Traction
- 4 ★
- added
- today
- Security policy
- no
SECURITY.mdat the repository root (which does not mean it is unsafe)
More like this
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.