# pinchbench

> pinchbench — olearycrew-pinchbench. Use this tool when you need to evaluate the performance of OpenClaw agents on real-world tasks, as it runs PinchBench benchmarks to assess their capabilities. It takes OpenClaw agent configurations as input and outputs performance metrics, providing valuable insights into agent effectiveness. Ideal for use cases where agent performance optimization is crucial, such as fine-tuning agent parameters or comparing agent performance across different tasks.

Canonical page: https://skillsregistry.net/skills/olearycrew-pinchbench  
JSON: https://api.skillsregistry.net/v1/skills/olearycrew-pinchbench

## Description

Run PinchBench benchmarks to evaluate OpenClaw agent performance across real-world tasks.

## Trust

- **Trust score (0–1):** 0.65
- **Verification tier:** scanned
- **Last scanned:** 2026-05-18

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** instructions
- **Runtime environment:** llm
- **Category:** productivity
- **Updated:** 2026-09-13

## Source

- **Source listing:** [ClawHub](https://clawskills.sh/skills/olearycrew-pinchbench)

## 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": "olearycrew-pinchbench"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/olearycrew-pinchbench` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/olearycrew-pinchbench/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