AgentShield
AgentShield is a AI agent skill for devops and ci/cd, published by affaan-m.
Scans Claude Code and agent setups for security vulnerabilities including hardcoded secrets, permission misconfigurations, hook injection, MCP server risks, and prompt injection.
Is it any good?
Effectiveness · benchmark pending
We haven't benchmarked AgentShield yet, so its score is provisional — computed from safety, maintenance, and documentation only, with popularity given no weight. When we benchmark it, we measure whether it produces materially better output than a frontier model alone, and that result leads the score. How we measure.
About AgentShield
Description · AI-summarised from the affaan-m README
AgentShield is a security auditing tool that scans AI agent configurations (particularly Claude Code setups) for 102 security rules across five categories: secrets detection, permission audits, hook analysis, MCP server security, and agent config review. It detects hardcoded API keys, credential leaks, overly permissive tool permissions, command injection vectors, data exfiltration risks, supply-chain vulnerabilities in MCP servers, and prompt injection surfaces. Available as a CLI tool, npm package, GitHub Action, and GitHub App integration, with support for multiple output formats (human-readable, JSON, HTML, evidence packs). Includes advanced features like Opus 4.6 adversarial analysis, auto-fixing of safe issues, runtime-confidence labeling to distinguish active configs from templates/examples, and a comprehensive false-positive audit workflow.
Where it fits
Browse related skills
Install AgentShield
Install
Install via agent tool
# install via agent tool
# see vendor docsInstall with your AI
Zero install · paste into ChatGPT, Claude, Cursor…
Install the "AgentShield" skill into my project by following all instructions at https://skillsdirectory.co/s/affaan-m-agentshield
Your assistant reads /s/affaan-m-agentshield and follows it — fetching the files from the source, placing them, and confirming before it runs anything. No account needed. How this works.
Frequently asked
FAQ
What does AgentShield do?
Scans Claude Code and agent setups for security vulnerabilities including hardcoded secrets, permission misconfigurations, hook injection, MCP server risks, and prompt injection.
How do I install AgentShield?
AgentShield is a AI agent skill. Primary install path: # install via agent tool # see vendor docs
What is the Skill Score for AgentShield?
AgentShield has a Skill Score of 6.8/10 (Solid). It is currently a provisional score — AgentShield hasn't been benchmarked yet, so it weights safety 75/100 (40%), maintenance 66/100 (30%), and documentation 71/100 (30%). Popularity carries zero weight until effectiveness is measured. See methodology.
Is AgentShield free?
AgentShield is published under the MIT license. See the vendor's source repository for any usage limits or paid tiers.
Related skills in DevOps and CI/CD
Related
Cloudflare MCP Server
Cloudflare's official MCP server — manage Workers, KV, R2, DNS, Zero Trust, and Cloudflare's wider platform from any MCP client.
LlamaDeploy
Async-first framework for deploying agentic LlamaIndex Workflows as production microservices with built-in scaling and queues.
XcodeBuildMCP
A Model Context Protocol server and CLI providing tools for AI agents to build, test, and manage iOS and macOS projects in Xcode.
Embed this score
For maintainers · always-current, links back
Put it in your README — it updates as the score changes.[](https://skillsdirectory.co/skills/affaan-m-agentshield)
<a href="https://skillsdirectory.co/skills/affaan-m-agentshield"><img src="https://skillsdirectory.co/badge/affaan-m-agentshield.svg" alt="Scored by skillsdirectory" height="28" /></a>