guards-for-agents

Name: guards-for-agents
Author: gramatr

gramatr ★ 0

Standalone security skill that deploys eight hooks and an MCP server to detect and redact secrets, PII, and prompt injections in Claude Code sessions, blocking credential leaks in tool output, agent messages, MCP results, and outbound publishes. Integrates with a guards-unified container service for detection logic, using gitleaks for local secret scanning and optional Hugging Face integration for injection detection.

Security DevOps 💲 Hugging Face No API key detected runs shell agent skill

Install

> npx skills add gramatr/claude-plugin-guards/tree/HEAD/guards

Installs via the community skills CLI (works with Claude Code and other agents).

Source: https://github.com/gramatr/claude-plugin-guards/tree/HEAD/guards

What it's made of

1 skill · 1 MCP server · 8 hooks

Commands: 0
Agents: 0
Skills: 1
MCP servers: 1
Hooks: 8 · PostToolUse, PreToolUse, SessionStart, UserPromptSubmit

What it needs & plugs into

API keys: none
Paid services: Hugging Face
External tools: bash uv
Talks to: nothing external detected

Analyzed 2026-10-25. Facts extracted from the plugin's files. Prose generated by claude-haiku-4-5-20251001.

Is this your plugin?

Claim it to keep the card accurate and enter the weekly contest. Requires signing in as the GitHub owner (gramatr).