# AgentShield

> Use this tool when you need to protect large language model (LLM) agents from malicious inputs and potential jailbreaks. AgentShield detects prompt-injection attacks at runtime, providing a secure gateway with high accuracy (F1 0.921) and low latency (2.4 ms). Ideal for applications requiring robust LLM security, AgentShield ensures reliable and trustworthy agent interactions.

Canonical page: https://skillsregistry.net/skills/io-github-dl-eigenart-agentshield-mcp  
JSON: https://api.skillsregistry.net/v1/skills/io-github-dl-eigenart-agentshield-mcp

## Description

Runtime gateway detecting prompt-injection and jailbreak for LLM agents. 2.4 ms, F1 0.921.

## Trust

- **Trust score (0–1):** 0.73
- **Verification tier:** scanned
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 0.1.1
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** ai-ml
- **Updated:** 2026-09-28

## Source

- **Source listing:** [MCP Registry](https://registry.modelcontextprotocol.io/v0/servers/io.github.dl-eigenart%2Fagentshield-mcp)
- **Repository:** <https://github.com/dl-eigenart/agentshield-platform>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "io-github-dl-eigenart-agentshield-mcp"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/io-github-dl-eigenart-agentshield-mcp` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/io-github-dl-eigenart-agentshield-mcp/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt