# SafeAgent

> Use this tool when you need to prevent duplicate actions and ensure idempotent execution of AI agent tasks, solving problems related to retry mechanisms and repeated tool calls. SafeAgent acts as an execution guard, taking in task inputs and outputting a guarded response to prevent redundant operations. Ideal for use cases where retrying failed tasks could lead to unintended consequences or errors.

Canonical page: https://skillsregistry.net/skills/io-github-azender1-safeagent  
JSON: https://api.skillsregistry.net/v1/skills/io-github-azender1-safeagent

## Description

Execution guard for AI agents — prevents duplicate tool calls on retry.

## Trust

- **Trust score (0–1):** 0.77
- **Verification tier:** verified
- **Last scanned:** 2026-09-19

## Facts

- **Version:** 0.1.14
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** ai-ml
- **Updated:** 2026-09-19

## Source

- **Source listing:** [MCP Registry](https://registry.modelcontextprotocol.io/v0/servers/io.github.azender1%2Fsafeagent)
- **Repository:** <https://github.com/azender1/SafeAgent>

## 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-azender1-safeagent"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/io-github-azender1-safeagent` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/io-github-azender1-safeagent/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