An offline scanner your agent runs before you trust a repo, skill, plugin or MCP server, and to audit what you already installed.
Pack · 2 skillsby alexgreensh
4 images
About
Repo Forensics is Alex Greenshpun's local security scanner for the AI-agent ecosystem, packaged as Agent Skills. Point your agent at a repository, skill, plugin or MCP server and it runs a set of scanners, correlates what they find, and returns a severity-ranked verdict before that code touches your machine. What its skills teach your agent - repo-forensics: audit a repo, skill, plugin or MCP server for prompt injection, MCP tool poisoning, hidden Unicode, credential exposure and exfiltration paths, malicious install hooks, known-malicious packages, CVEs and CISA KEV exposure, and traces of past supply-chain attacks - forensify: audit the agent stack you already have (installed skills and plugins, MCP server configs, hooks, memory and settings files, and credential file permissions) across Claude Code, Codex, OpenClaw and NanoClaw - Correlate findings into compound threats (an environment read plus an outbound POST becomes data exfiltration) and sort them into BLOCK, WARN and INFO tiers - Report as text, JSON or SARIF for the GitHub Security tab, with exit codes for CI How it works - It's pure Python with no dependencies, and scanning runs on your machine. The only network lookups are package names and versions (for CVEs and signatures) and links the scanned repo points to; --offline turns those off. - Add the skills to your agent's skills folder; the maker documents Claude Code, Codex, OpenClaw, NanoClaw and Cursor. For automatic scanning of every clone and install, the maker's Claude Code and Codex plugin adds hooks on top. - Forensify is read-only: it doesn't fix or quarantine anything, and it reads credential file metadata, never the values. Good to know - Licensed PolyForm Noncommercial 1.0.0: free for personal, research and education use and for very small businesses under the maker's written permission; larger commercial use needs a license from the maker. - A community project by Alex Greenshpun (alexgreensh).
Once it’s added
“Audit this repo before I add it as a dependency”
“Is this MCP server safe to use?”
“Run forensics on ~/Downloads/new-plugin”
“Audit every skill, plugin and MCP server I have installed”
“Check this project's .claude settings, hooks and .mcp.json for anything suspicious”
“Give me the findings as SARIF for GitHub”
Before you add it
Reads only
The repo, skill, plugin or MCP server you point it at · Your installed agent stack: skills, plugins, MCP configs, hooks and settings (file metadata only for credentials) · Package names and versions, to look up CVEs and known-malicious packages (off with --offline)
Makes changes · your agent can ask first
Writes a severity-ranked report as text, JSON or SARIF
Pack · 2 skills, each with its own page
From people who use it
Step by step
Nothing new reaches your agent until a person has looked at it.
No reviews yet. Used Repo Forensics? Say how it went.
Write a reviewRuns on your machine.
Runs on your machine.
Runs on your machine.
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.copilot/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.copilot/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.cursor/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.cursor/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.claude/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.claude/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.agents/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.agents/skills/repo-forensics
Runs on your machine.
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.gemini/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.gemini/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify .agents/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics .agents/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.openclaw/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.openclaw/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.config/opencode/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.config/opencode/skills/repo-forensics
Runs on your machine.
npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/forensify ~/.hermes/skills/forensify npx -y degit alexgreensh/repo-forensics/plugins/repo-forensics/skills/repo-forensics ~/.hermes/skills/repo-forensics
Scanned and passed · checked 2 h ago
Nobody has shared how they use Repo Forensics yet.
Skills people install alongside Repo Forensics.
Scanned on every change, and again when our rules improve.
GeminiSupported
GitHub CopilotSupportedTell us and a person on the trust team will look at it. Or email [email protected].
Skills people install alongside Repo Forensics.
Trail of Bits' skills for fuzzing and crypto testing, from its Testing Handbook: libFuzzer, AFL++, cargo-fuzz, Atheris, OSS-Fuzz and more.
Lets your agent find vulnerabilities in your code, prove they're real against your running app, and fix them: SCA, secrets, SAST and DAST.
Has your agent diff two Windows builds with WinDiff and explain what changed in syscalls, symbols and mitigations, and why it matters.
Security scan
Passed
Safe to install in every agent it supports. Runs in your agent, as instructions.
