squirrelscan
Official squirrelscan plugins: website QA for coding agents
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Check it deeper on your machine
It has changed since its scan on 11 Nov 2026.
As a Claude Code plugin, install it and then ask Claude Code:
/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/squirrelscan/skills with scanskillsafety
Or from a terminal, clone it and run it:
git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/squirrelscan/skills'
It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.
Badge for your README
Markdown
[](/listings/squirrelscan--skills)HTML
<a href="/listings/squirrelscan--skills"><img src="/api/listings/squirrelscan--skills/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Pick the path for your tool:
| Tool | Skills only | Skills + MCP server | |------|-------------|---------------------| | Claude Code | `npx skills add squirrelscan/skills` | `/plugin marketplace add squirrelscan/skills` then `/plugin install squirrelscan@squirrelscan` | | Cursor | `npx skills add squirrelscan/skills` | the Cursor plugin in this repo, or the [one-click MCP link](cursor://anysphere.cursor-deeplink/mcp/install?name=squirrelscan&config=eyJ1cmwiOiJodHRwczovL21jcC5zcXVpcnJlbHNjYW4uY29tL21jcCJ9) plus skills | | OpenAI Codex | `npx skills add squirrelscan/skills` (lands in `.agents/skills/`) | skills plus the MCP server in `~/.codex/config.toml` ([docs](https://docs.squirrelscan.com/developers/agents/codex)) | | Gemini CLI, GitHub Copilot, Amp and other Agent Skills tools | `npx skills add squirrelscan/skills` | skills plus the MCP server ([docs](https://docs.squirrelscan.com/developers/agents)) | | Any Agent Plugins client | | this repo: `plugin.json`, `mcp.json` and `skills/` | | squirrel CLI | `squirrel skills install` | | | Manual | clone + symlink `skills/*` into your agent's skills dir | |
Agent Skills via npx (works everywhere)
npx skills add squirrelscan/skills
Installs both skills (`squirrelscan` and `audit-website`) for whichever agents you select: Claude Code, Codex, Cursor, Gemini CLI, Amp, and more. Skills follow the [Agent Skills](https://agentskills.io) standard, so the same `SKILL.md` works across tools. To install just one skill:
npx skills add squirrelscan/skills --skill audit-website
Claude Code plugin (recommended for Claude Code)
/plugin marketplace add squirrelscan/skills /plugin install squirrelscan@squirrelscan
One step installs both skills and connects the hosted squirrelscan MCP server.
Claude Code channel (opt in)
A separate plugin, `squirrelscan-channel`, runs `squirrel channel` and pushes cloud audit events into your running session, so your agent can pull the report and fix what it finds. It is not part of the main plugin, so nobody gets a background process they did not ask for.
/plugin install squirrelscan-channel@squirrelscan claude --channels plugin:squirrelscan-channel@squirrelscan
It needs a CLI release with the `channel` command, and until squirrelscan is on Anthropic's channel allowlist you need the development flag instead of `--channels`. Details: [plugins/squirrelscan-channel](plugins/squirrelscan-channel/README.md).
Cursor
1. **Skills**: `npx skills add squirrelscan/skills`. Cursor reads Agent Skills from `.cursor/skills/`, `.agents/skills/`, and their `~/` equivalents (it also picks up `~/.claude/skills/`). 2. **MCP only, one click**: [Add squirrelscan MCP to Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=squirrelscan&config=eyJ1cmwiOiJodHRwczovL21jcC5zcXVpcnJlbHNjYW4uY29tL21jcCJ9) 3. **Plugin** (skills + MCP): this repo is a Cursor plugin (`.cursor-plugin/`). Teams can import it as a team marketplace (Dashboard, **Plugins & MCPs**, **Add Marketplace**, **Import from Repo**), or clone it into `~/.cursor/plugins/local/squirrelscan` and reload the window ([Cursor docs](https://cursor.com/docs/plugins#test-plugins-locally)).
OpenAI Codex
Codex reads skills from `.agents/skills/` (project) or `~/.agents/skills/` (global). `npx skills add squirrelscan/skills` installs there, or clone and symlink the `skills/*` directories.
From the squirrel CLI
squirrel skills # what is installed where, against the latest version squirrel skills install squirrel skills update
From v0.0.99 the CLI installs the skills itself, from `manifest.json`, into `~/.claude/skills` and `~/.agents/skills` (or a repo's own with `--project`), takes over installs made with `npx skills`, and keeps them current with its own auto-update. See [docs.squirrelscan.com/cli/skills](https://docs.squirrelscan.com/cli/skills).
Manual
git clone https://github.com/squirrelscan/skills.git
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 24 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.
- {
- "name": "squirrelscan",
- "owner": {
- "name": "squirrelscan",
- "email": "support@squirrelscan.com"
- },
- "description": "Official squirrelscan plugins: website QA for coding agents",
- "plugins": [
- {
- "name": "squirrelscan",
- "source": "./",
- "description": "Website QA tool built for coding agents (Claude Code, Cursor). 300+ rules across SEO, performance, security, accessibility & agent experience: audit from the CLI, fix findings in code, publish reports, and connect over MCP.",
- "category": "productivity",
- "tags": ["seo", "audit", "website", "performance", "security"]
- },
- {
- "name": "squirrelscan-channel",
- "source": "./plugins/squirrelscan-channel",
- "description": "Opt-in Claude Code channel: pushes squirrelscan cloud audit events into your running session so your agent can pull the report and fix what it finds. Enable with --channels plugin:squirrelscan-channel@squirrelscan.",
- "category": "productivity",
- "tags": ["audit", "website", "channel", "notifications"]
- }
- ]
- }
README.md · README · 196 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.
- 
- # squirrelscan skills and plugins
- **Website audits for AI agents: skills, plugins, and MCP in one repo**
- ## What is squirrelscan?
- [squirrelscan](https://squirrelscan.com) is a website audit tool built for AI agents. It crawls a site, analyzes every page against 300+ rules across SEO, performance, security, accessibility, content, and more, and returns a health score plus concrete, fixable issues.
- **Features:**
- - 300+ audit rules across 22 categories
- - Leaked secrets detection (96 patterns: OpenAI, Anthropic, AWS, Stripe, and more)
- - Multiple output formats: console, text, json, markdown, llm, html
- - Diff reports for regressions between audits
- - LLM-native output for AI-assisted debugging and optimization
- - Optimized for CI/CD pipelines and automation
- ## What's in this repo
- This repo is the one home of the squirrelscan agent skills and agent plugins. Every install path below reads it; nothing else carries a copy.
- | Skill | What it does |
- |-------|--------------|
- | [`squirrelscan`](skills/squirrelscan/SKILL.md) | Operate the CLI: install, login, run audits, publish reports, credits, API keys, the entity map, MCP setup, config, troubleshooting |
- | [`audit-website`](skills/audit-website/SKILL.md) | The fix loop: audit a site, map findings to code, fix in batches, re-audit until it scores well |
- Each plugin bundles both skills and the hosted squirrelscan MCP server:
- | Plugin | Files |
- |--------|-------|
- | Claude Code marketplace + plugin | `.claude-plugin/marketplace.json`, `.claude-plugin/plugin.json`, `.mcp.json` |
- | Claude Code channel plugin (opt in) | `plugins/squirrelscan-channel/` |
- | Cursor plugin | `.cursor-plugin/plugin.json`, `.cursor-plugin/mcp.json` |
- | [Agent Plugins](https://agent-plugins.org) 1.0.0 (open standard, loads in Cursor) | `plugin.json`, `mcp.json` |
- Codex reads each skill's `agents/openai.yaml` for its display metadata.
- `manifest.json` lists every skill's version and every file in it with its sha256 and size. The squirrel CLI installs and updates the skills from it: it downloads only the files whose hash changed and verifies each one before writing anything.
- ## Prerequisites
- All skills drive the **squirrel CLI**, which must be installed and in PATH.
- **Install:** [squirrelscan.com/download](https://squirrelscan.com/download)
- **Verify:**
- ```bash
- squirrel --version
- ```
- ## Installing
- Pick the path for your tool:
- | Tool | Skills only | Skills + MCP server |
- |------|-------------|---------------------|
- | Claude Code | `npx skills add squirrelscan/skills` | `/plugin marketplace add squirrelscan/skills` then `/plugin install squirrelscan@squirrelscan` |
- | Cursor | `npx skills add squirrelscan/skills` | the Cursor plugin in this repo, or the [one-click MCP link](cursor://anysphere.cursor-deeplink/mcp/install?name=squirrelscan&config=eyJ1cmwiOiJodHRwczovL21jcC5zcXVpcnJlbHNjYW4uY29tL21jcCJ9) plus skills |
- | OpenAI Codex | `npx skills add squirrelscan/skills` (lands in `.agents/skills/`) | skills plus the MCP server in `~/.codex/config.toml` ([docs](https://docs.squirrelscan.com/developers/agents/codex)) |
- | Gemini CLI, GitHub Copilot, Amp and other Agent Skills tools | `npx skills add squirrelscan/skills` | skills plus the MCP server ([docs](https://docs.squirrelscan.com/developers/agents)) |
- | Any Agent Plugins client | | this repo: `plugin.json`, `mcp.json` and `skills/` |
- | squirrel CLI | `squirrel skills install` | |
- | Manual | clone + symlink `skills/*` into your agent's skills dir | |
- ### Agent Skills via npx (works everywhere)
- ```bash
- npx skills add squirrelscan/skills
- ```
- Installs both skills (`squirrelscan` and `audit-website`) for whichever agents you select: Claude Code, Codex, Cursor, Gemini CLI, Amp, and more. Skills follow the [Agent Skills](https://agentskills.io) standard, so the same `SKILL.md` works across tools. To install just one skill:
- ```bash
- npx skills add squirrelscan/skills --skill audit-website
- ```
- ### Claude Code plugin (recommended for Claude Code)
- ```
- /plugin marketplace add squirrelscan/skills
- /plugin install squirrelscan@squirrelscan
- ```
- One step installs both skills and connects the hosted squirrelscan MCP server.
- ### Claude Code channel (opt in)
- A separate plugin, `squirrelscan-channel`, runs `squirrel channel` and pushes cloud audit events into your running session, so your agent can pull the report and fix what it finds. It is not part of the main plugin, so nobody gets a background process they did not ask for.
- ```
- /plugin install squirrelscan-channel@squirrelscan
- claude --channels plugin:squirrelscan-channel@squirrelscan
- ```
- It needs a CLI release with the `channel` command, and until squirrelscan is on Anthropic's channel allowlist you need the development flag instead of `--channels`. Details: [plugins/squirrelscan-channel](plugins/squirrelscan-channel/README.md).
- ### Cursor
- 1. **Skills**: `npx skills add squirrelscan/skills`. Cursor reads Agent Skills from `.cursor/skills/`, `.agents/skills/`, and their `~/` equivalents (it also picks up `~/.claude/skills/`).
- 2. **MCP only, one click**: [Add squirrelscan MCP to Cursor](cursor://anysphere.cursor-deeplink/mcp/install?name=squirrelscan&config=eyJ1cmwiOiJodHRwczovL21jcC5zcXVpcnJlbHNjYW4uY29tL21jcCJ9)
- 3. **Plugin** (skills + MCP): this repo is a Cursor plugin (`.cursor-plugin/`). Teams can import it as a team marketplace (Dashboard, **Plugins & MCPs**, **Add Marketplace**, **Import from Repo**), or clone it into `~/.cursor/plugins/local/squirrelscan` and reload the window ([Cursor docs](https://cursor.com/docs/plugins#test-plugins-locally)).
- ### OpenAI Codex
- Codex reads skills from `.agents/skills/` (project) or `~/.agents/skills/` (global). `npx skills add squirrelscan/skills` installs there, or clone and symlink the `skills/*` directories.
- ### From the squirrel CLI
- ```bash
- squirrel skills # what is installed where, against the latest version
- squirrel skills install
- squirrel skills update
- ```
- From v0.0.99 the CLI installs the skills itself, from `manifest.json`, into `~/.claude/skills` and `~/.agents/skills` (or a repo's own with `--project`), takes over installs made with `npx skills`, and keeps them current with its own auto-update. See [docs.squirrelscan.com/cli/skills](https://docs.squirrelscan.com/cli/skills).
- ### Manual
- ```bash
- git clone https://github.com/squirrelscan/skills.git
- ```
- Then copy or symlink `skills/squirrelscan` and `skills/audit-website` into your agent's skills directory (`make link` does this for Claude Code and `.agents/skills` consumers).
- ## Updates and versions
- - **Skills** carry their own version in `SKILL.md` (`metadata.version`). Bump it with every content change.
- - **Plugins** carry a semver `version` (`claude plugin validate --strict` and the OpenAI listing both need one). The `squirrelscan` plugin's version is the squirrelscan skill's `metadata.version` padded to three parts (skill `1.16` is plugin `1.16.0`) in `.claude-plugin/plugin.json`, `.cursor-plugin/plugin.json` and `plugin.json`; marketplace entries carry none. Claude Code treats the version string as the update key, so an install updates only when we bump it: **bump the skill version with every skill content change, and the plugin versions move with it** (CI fails on drift). Changes that touch only plugin files (MCP config, manifests) need a skill version bump too, or installs will not see them. `squirrelscan-channel` has its own semver, bumped when that plugin changes. Third-party marketplaces don't auto-update by default: run `/plugin marketplace update squirrelscan` then `/plugin update squirrelscan@squirrelscan`, or turn on auto-update for the marketplace in `/plugin` ([Claude Code docs](https://code.claude.com/docs/en/plugins-reference#version-management)).
- - **npx installs** update with `npx skills update`, `squirrel skills update`, or the squirrel CLI's auto-update.
- `bun run scripts/check-skills.ts` (CI runs it on pull requests and on pushes to `main`) checks that every `SKILL.md` parses the way installers read it, and that the plugin manifests agree: same name, same description, a version that matches the skill, one MCP endpoint, and a complete OpenAI listing block.
- ## Moving from squirrelscan/squirrelscan
- The skills and plugins used to ship from the [squirrelscan/squirrelscan](https://github.com/squirrelscan/squirrelscan) repo too. They now ship only from here.
- - **Claude Code**: the old `squirrelscan/squirrelscan` marketplace now points its plugin at this repo, so it keeps working. Pick up the move with `/plugin marketplace update squirrelscan` then `/plugin update squirrelscan@squirrelscan`. To switch marketplaces instead, remove the old one first: both are named `squirrelscan`, and Claude Code won't add a second marketplace under a name that is already taken. Removing it also uninstalls the plugin, so reinstall it from here:
- ```
- /plugin marketplace remove squirrelscan
- /plugin marketplace add squirrelscan/skills
- /plugin install squirrelscan@squirrelscan
- ```
- - **npx skills**: from squirrel v0.0.99, `squirrel skills install` (or `squirrel skills update`) takes over installs made with `npx skills`, whichever repo they came from, and removes them from its lock file. Without the CLI: `npx skills add squirrelscan/skills -g`.
- - **Cursor plugin**: install it from this repo as above. Cursor's marketplace manifests can't point at another repository, so a copy of the old one does not follow the move.
- ## MCP server
- The hosted MCP server lives at `https://mcp.squirrelscan.com/mcp` (streamable-http; OAuth or `Authorization: Bearer sq_...` API key). A local stdio server is available via `squirrel mcp`. Docs: [docs.squirrelscan.com/developers/mcp](https://docs.squirrelscan.com/developers/mcp)
- ## Example prompts
- ```
- Audit this website and fix all errors and warnings
- ```
- ```
- Run an audit on example.com and show me the top 5 critical issues
- ```
- ```
- Check my site for broken links and leaked secrets
- ```
- ```
- Re-audit after the deploy and diff against the last report
- ```
- ## Contributing
- Skill and plugin changes land here, and only here. Contributions are welcome! To suggest new skills or improvements:
- 1. Open an issue to discuss your idea
- 2. Fork this repository
- 3. Create a feature branch
- 4. Submit a pull request
- After changing anything under `skills/`, regenerate the manifest and commit it with the change (CI fails when it is stale):
- ```bash
- bun run scripts/manifest.ts
- ```
- Bump the skill's `metadata.version` in its `SKILL.md` too, so installs can tell the versions apart. For `squirrelscan` also set the plugin `version` in the three manifests to the same number plus `.0` (`1.16` becomes `1.16.0`); `check-skills.ts` fails on a mismatch.
- All skills follow the [Agent Skills Standard](https://agentskills.io/specification).
- ## License
- MIT License. See [LICENSE](LICENSE) file for details.
- ---
- **Learn more:** [docs.squirrelscan.com](https://docs.squirrelscan.com) | [squirrelscan.com](https://squirrelscan.com)
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.