usmanmanii/realitygate / skills/scan
scan
VerifiedCheck whether an agent skill, Claude Code plugin, MCP config or AI coding extension is safe before installing it — scans for remote code execution, obfuscated payloads, hidden Unicode instructions, credential or .env access, data-leak endpoints, prompt injection, permission bypass, hooks, MCP servers and install scripts. Use when the user asks if a skill, plugin, extension or repository is safe, trustworthy or malware-free, or before installing one from the internet.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 1.9 KB |
SKILL.md
View on GitHub---
name: scan
description: Check whether an agent skill, Claude Code plugin, MCP config or AI coding extension is safe before installing it — scans for remote code execution, obfuscated payloads, hidden Unicode instructions, credential or .env access, data-leak endpoints, prompt injection, permission bypass, hooks, MCP servers and install scripts. Use when the user asks if a skill, plugin, extension or repository is safe, trustworthy or malware-free, or before installing one from the internet.
---
# RealityGate scan: check a skill or plugin before installing it
1. Get the package onto disk **without installing or running it**. For a GitHub repo, clone it
into a temporary folder: `git clone --depth 1 <url> <temp-folder>`. Never run its install
scripts, `npm install` or its hooks.
2. Scan it:
```bash
node "${CLAUDE_PLUGIN_ROOT}/bin/realitygate.mjs" scan-skills <folder>
```
(Outside the plugin: `npx --yes realitygate scan-skills <folder>`.)
3. Report to the user in plain language:
- **HIGH** findings: quote each flagged line and explain what it would do. Recommend not
installing until a human has read every one.
- **What it would run**: every hook (these run automatically on agent events), every MCP server
(a remote one can receive your data), install scripts, and files that start processes.
- **Excluded by the publisher**: the publisher chose not to have these files scanned. Read
them yourself, or scan again with those files included.
4. Always end with this sentence: a clean scan is not a guarantee. Malicious instructions can be
written in ways no pattern catches. Suggest reading the `SKILL.md` files, checking who the
publisher is, and pinning a specific version or commit.
Rules:
- Never execute anything from the scanned package while investigating it.
- Don't summarise HIGH findings away. Show the exact line.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |