# llm-grace

> llm-grace — hyperpolymath-llm-grace. Use this tool when you need to ensure reliable and fault-tolerant operation of concurrent LLM/agent terminals, solving problems of catastrophic failures and data loss. It provides a shared crash-safe ledger, legible post-mortems, and cheap load-shedding, accepting input from LLM/agent terminals and outputting diagnostic information and failure reports. Ideal for use in high-availability and mission-critical applications where graceful degradation is essential.

Canonical page: https://skillsregistry.net/skills/hyperpolymath-llm-grace  
JSON: https://api.skillsregistry.net/v1/skills/hyperpolymath-llm-grace

## Description

Graceful degradation for concurrent LLM/agent terminals: cheap /proc-signature load-shedding, a shared crash-safe ledger, and legible post-mortems. Fail legibly, not catastrophically.

## Trust

- **Trust score (0–1):** 1.00
- **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/hyperpolymath/llm-grace)

## 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": "hyperpolymath-llm-grace"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/hyperpolymath-llm-grace` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/hyperpolymath-llm-grace/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