Plugin Auditor — Security skill for Claude Code
Statically audit a Claude Code plugin bundle for dangerous or injected components (hooks, bin/, monitors, workflows, scripts, skill/command/agent injection, manifest provenance) before enabling it — a.
How to install Plugin Auditor
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Claude-Booster/plugin-auditor and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Plugin Auditor does
Statically audit a Claude Code plugin bundle for dangerous or injected components (hooks, bin/, monitors, workflows, scripts, skill/command/agent injection, manifest provenance) before enabling it — a self-contained Claude Code skill.
Alternatives in Security
- Agentseal — Security toolkit for AI agents 344 ★
- Openclaw Ops — OpenClaw operations skill with health checks, repair scripts, watchdogs, update triage, and security scans 225 ★
- Claude Skill Antivirus — Security scanner for Claude Code Skills — 9 engines detect malicious patterns, data exfiltration, dangerous op 64 ★
README
-56c596?style=flat-square&labelColor=0c1a28){kind=icon}
What it is
A Claude Code **plugin is a filesystem bundle**, not a single MCP server — and when you enable it, many of its parts **execute** (hooks, `bin/` binaries, background monitors, `workflows/*.js`, LSP servers, bundled scripts) or **load into the model's context** (the markdown of its skills, commands, and agents). Reviewing the plugin's *code* misses most of this.
**plugin-auditor** statically inspects a plugin bundle and flags the risky parts in plain English. It is **read-only**: it never runs any hook, script, binary, workflow, monitor, or MCP server from the plugin, and never prints secret values (key names only).
It sits alongside the sibling tools: **config-guard** gates/quarantines untrusted config; **plugin-auditor** deep-analyzes a plugin bundle; **mcp-auditor** audits the MCP tool metadata within it — plugin-auditor delegates to it.
What it inspects
- Hooks — commands that auto-run on Claude events (flags
curl|sh, credential reads). bin/— executables placed on the Bash tool's PATH (Claude can run them
Related Skills
Trust Me Bro
Every skill you install says trust me bro. This one reads it first. Pre-install audit for agent skills, plugin
Agent Security Hooks
Security hooks for AI coding assistants (Claude Code, Cursor, Gemini CLI) - block dangerous commands, protect
AegisGate
Open-source security gateway for LLM APIs — prompt injection detection, PII redaction, dangerous response sani
Kiteguard
Open-source runtime security guardrails for Claude Code, Cursor, and Gemini CLI — blocks dangerous commands, d
DeepSafe Scan — Preflight Security Scanner For AI Coding Agents
Full-featured preflight security scanner across 5 dimensions: Posture (config), Skill (skills & MCP), Memory (
Claude Guardrails
Hardened security configuration for Claude Code; permission deny rules, shell hooks, and prompt injection defe
Related Agents
Laya Vetter
Security-vets a candidate plugin, skill or MCP server repo before installation (hooks, install scripts, creden
Redteam
Use for anything under redteam/ — building prompt-injection and exfiltration suites, mapping what an injected
Pack Coder
Use to execute approved implementation plans against pack source — writing/editing scripts, fixtures, configs,
