whereismyai
AgentGuard
AgentGuard is an AI agent skill scanner that checks skill files for security risks and hidden capabilities before installation.
AgentGuard is an AI agent skill scanner designed to evaluate SKILL.md files for security risks, hidden capabilities, and quality issues before they touch your codebase. Users paste skill instructions into the scanner to receive an actionable trust report in seconds. The tool combines deterministic pattern checks with AI-assisted analysis to surface potential dangers such as shell commands, file system access, environment variables, network requests, external downloads, credential access, suspicious URLs, obfuscated instructions, prompt injection patterns, and destructive operations. It supports inputs up to 80,000 characters. The platform is free to use and requires no account to get started.
AgentGuard at a glance
Key features
- Scan SKILL.md files up to 80,000 characters
- Detect shell command execution instructions
- Identify file system and project file access
- Flag external network requests and URLs
- Surface environment variable and credential access
- Catch obfuscated instructions and prompt injections
- Generate security and quality scores
Use it for
- Scan AI agent skills for security risks before installation
- Review shell commands and network requests in skill files
- Analyze prompt injection patterns in agent instructions
Alternatives to AgentGuard
- Cisco Skill Scanner — Combines pattern matching, dataflow analysis, and an optional LLM judge for threat detection.
- Agent Skills — Encodes senior-engineer workflows and quality gates into 24 skills.
- Tech Leads Agent Skills — Extends AI coding agents with security-scanned, specialized workflows and tools.
- SkillTotal — Scans code locally without accounts, telemetry, API tokens or executing the files.
- Stop That Shit — Enforces task boundaries to stop unnecessary AI agent actions.
- Security Audit Skill — Automated multi-stage security auditing with independent verification.
- Secret Guard — Prevents sensitive credentials from reaching third-party AI vendor servers.
- Repo Guards — It prevents AI agents from performing destructive actions across multiple repositories.
- Laya Guard — Judges every tool call locally on your laptop in ~0.1s with no per-decision billing.
- Prediction Guard — Enforces NIST and OWASP policies on every model and agent from inside your security boundary.
- ARTEX — Runs autonomous multi-agent penetration testing with a dual-graph architecture.
- AgentGate — Blocks attacks in under 35ms with a local neural classification head.
All AgentGuard alternatives