> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scanaislop.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Give aislop to your coding agent

> Copy one prompt or URL so Claude Code, Cursor, Codex, and other agents install the aislop skill, hooks, and AGENTS.md standard.

Paste a prompt, a URL, or three commands. The agent then installs the [aislop skill](/integrations/agent-skill), [hooks](/cli/hooks), and repo instructions so every edit is held to the same quality gate.

## Paste this into any agent

```text theme={null}
Set up aislop as the quality gate for this repository.

Read https://docs.scanaislop.com/agents
and https://scanaislop.com/.well-known/agent-skills/aislop/SKILL.md

Then:
1. npx skills add scanaislop/skills
2. npx aislop@latest hook install
3. Add the AGENTS.md template from that docs page (also CLAUDE.md / GEMINI.md if those files exist)
4. npx aislop@latest scan
5. npx aislop@latest fix

Do not disable rules. Scan after edits. Fix mechanical findings with aislop fix. Fix the rest in-session. Re-scan before you say done.
```

That block is enough. The agent can fetch the pages, install the skill and hooks, and write the standard.

## Or send a URL

Give the agent this page:

```text theme={null}
https://docs.scanaislop.com/agents
```

Or the loadable skill file:

```text theme={null}
https://scanaislop.com/.well-known/agent-skills/aislop/SKILL.md
```

Related machine-readable surfaces:

| URL                                                                                                              | What it is                                   |
| ---------------------------------------------------------------------------------------------------------------- | -------------------------------------------- |
| [scanaislop.com/llms.txt](https://scanaislop.com/llms.txt)                                                       | Compact product and command index for agents |
| [scanaislop.com/.well-known/agent-skills/index.json](https://scanaislop.com/.well-known/agent-skills/index.json) | Skill discovery index                        |
| [skills.sh/scanaislop/skills](https://www.skills.sh/scanaislop/skills)                                           | Public skill listing                         |

## Or run the commands yourself

```bash theme={null}
npx skills add scanaislop/skills
npx aislop@latest hook install
npx aislop@latest scan
```

`npx skills add scanaislop/skills` installs the aislop skill into the coding agents on this machine. `aislop hook install` wires a scan after every agent edit (Claude Code, Cursor, Gemini CLI, and others). See [Agent skill](/integrations/agent-skill) and [Hooks](/cli/hooks) for flags.

Add `--quality-gate` when you want the agent blocked if the score drops:

```bash theme={null}
npx aislop@latest hook install --claude --quality-gate
npx aislop@latest hook baseline
```

## AGENTS.md template

Drop this at the repo root as `AGENTS.md`. If the repo already has `CLAUDE.md` or `GEMINI.md`, add `@AGENTS.md` on its own line (or copy the same body).

```md theme={null}
# aislop

aislop is the quality gate for AI generated code in this repo. Scan after you edit. Treat findings as blocking, not advisory.

## Install (once)

    npx skills add scanaislop/skills
    npx aislop@latest hook install

## On every edit

1. npx aislop@latest scan --changes --json (or --staged before commit)
2. npx aislop@latest fix for mechanical issues (imports, comments, formatting)
3. Fix remaining findings in-session in this project's style
4. Re-scan. Do not say done with unresolved errors or fixable warnings

## Do not

- Disable rules or add blanket aislop-ignore comments to pass
- Dump raw JSON at the user; triage findings
- Run aislop fix -f unless asked (it can rewrite manifests)

Config lives in `.aislop/config.yml`. Architecture rules, if any, live in `.aislop/rules.yml`.
```

`aislop hook install` also writes agent-specific instruction files (`AISLOP.md`, Cursor rules, Copilot instructions, and so on). Keep `AGENTS.md` as the portable copy.

## Handoff prompt (remaining findings)

After `aislop fix`, remaining issues need judgment. Print a prompt you can paste into any agent:

```bash theme={null}
npx aislop@latest fix --prompt
```

Or open a specific agent with that prompt already filled:

```bash theme={null}
npx aislop@latest fix --claude
npx aislop@latest fix --codex
npx aislop@latest fix --cursor
```

See [aislop agent](/cli/agent) for the local repair session, and [aislop fix](/cli/fix) for `--prompt` and agent flags.

## MCP tools

If the agent speaks MCP, add:

```json theme={null}
{
  "mcpServers": {
    "aislop": {
      "command": "npx",
      "args": ["-y", "aislop-mcp"]
    }
  }
}
```

That exposes `aislop_scan`, `aislop_fix`, `aislop_why`, and `aislop_baseline`. Details: [MCP server](/integrations/mcp-server).

## How the pieces stack

| Piece                              | Job                                                                        |
| ---------------------------------- | -------------------------------------------------------------------------- |
| [Skill](/integrations/agent-skill) | Teaches the agent when to scan, how to fix, and when it is allowed to stop |
| [Hooks](/cli/hooks)                | Runs a scan after every edit and feeds findings back the same turn         |
| `AGENTS.md`                        | Portable repo instructions every agent can load                            |
| [MCP](/integrations/mcp-server)    | Callable tools inside the editor                                           |
| [aislop agent](/cli/agent)         | Isolated local repair session with Codex, Claude Code, or OpenCode         |

Use all of them. The skill is the procedure. The hook is the enforcement. `AGENTS.md` is the written standard.