# agent-failure-archive

> agent-failure-archive — hanbeenmoon-agent-failure-archive. Use this tool when you need to analyze and learn from AI agent failures to improve system reliability and performance. The agent-failure-archive provides 186 production post-mortems from an 8-month multi-session AI agent system, offering valuable insights into common pitfalls and error patterns. It takes in system failure data as input and outputs actionable lessons and improvements, ideal for use in debugging and optimization contexts.

Canonical page: https://skillsregistry.net/skills/hanbeenmoon-agent-failure-archive  
JSON: https://api.skillsregistry.net/v1/skills/hanbeenmoon-agent-failure-archive

## Description

186 production post-mortems from an 8-month multi-session AI agent system, sold per call over x402

## Trust

- **Trust score (0–1):** 0.86
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **Updated:** 2026-09-28

## Source

- **Source listing:** [GitHub](https://github.com/HanbeenMoon/agent-failure-archive)

## 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": "hanbeenmoon-agent-failure-archive"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/hanbeenmoon-agent-failure-archive` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/hanbeenmoon-agent-failure-archive/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