Discover the toolingbehind modern intelligence
The operating system for AI infrastructure discovery. Search a live map of MCP servers, AI agents, LLM tools, automation systems, and developer infrastructure.
agentseal
Security toolkit for AI agents. Scan your machine for dangerous skills and MCP configs, monitor for supply chain attacks, test prompt injection resistance, and audit live MCP servers for tool poisoning.
281stars
38forks
Python
agent-securityai-agentai-agentsai-securityclillmllm-securitymcpmcp-securitymodelcontextprotocolpentestingprompt-injection
Quality signals
ActiveNo testsNo CIThin docsNo licensePopular
AI summary
Auto-generated overview of what this skill does and when to reach for it.
Open-source security scanner that red-teams AI agents by detecting prompt injections, MCP poisoning, and dangerous skill files on your machine.
Use this when
- You suspect an AI agent has been compromised by a supply chain attack
- You want to test a system prompt against adversarial attacks before deploying it
- You need to audit MCP servers for hidden tool descriptions that exfiltrate data
- You’re running a local LLM and want free, offline security testing
- You need continuous monitoring for new threats in agent config files
Don't use when
- You need runtime protection against already-running malicious agents
- You’re auditing non-agent software or general system security
- You require a cloud-based security service with managed dashboards
Capabilities
- Scans 28+ AI agents (Claude, Cursor, VS Code, etc.) for malicious skill files
- Detects prompt injection via 225+ adversarial probes with deterministic trust scores
- Audits live MCP servers for tool description poisoning without LLM calls
- Monitors agent configs in real-time and quarantines payloads automatically
- Tracks config changes with SHA-256 baselines to spot supply chain attacks
- Outputs SARIF/JSON for CI/CD pipelines and GitHub Security tab
InstallMCPRun via uvx (Python)
Universal installANY TOOL
$ npx skillcity@latest getagentseal/agentsealAuto-detects the skill or MCP server in this repo and installs it for Claude Code, Codex, Cursor, Cline, OpenCode, Gemini CLI, Anti-Gravity, OpenClaw and Hermes.
Requires Node.js 18+ (ships with npx).
Or install manually
uvx agentseal
MCP Configuration
Add this to your claude_desktop_config.json
{
"mcpServers": {
"agentseal": {
"command": "uvx",
"args": [
"agentseal"
]
}
}
}Loading README…
Ask a simulated agent to run this skill. It loads the real SKILL.md and reasons over your task with NVIDIA.
agentseal — sandboxAI sandbox
Pick a quick prompt below or type your own — watch the agent execute this skill step by step.
guest@skillcity:~$Try a prompt