Agent skill · research science · aiskillstore
pinchbench
Run PinchBench benchmarks to evaluate OpenClaw agent performance across real-world tasks. Use when testing model capabilities, comparing models, submitting benchmark results to the leaderboard, or checking how well your OpenClaw setup handles calendar, email, research, coding, and multi-step workflows.
Why this skill is useful
Provides specific benchmarking commands and task definitions for evaluating LLM performance that the AI wouldn't reliably generate on its own.
What it needs
Requires python, uv installed locally. About 2k tokens when loaded. Last updated 2026-08-07. 405 stars on the source repository.
What this skill does
PinchBench Benchmark Skill PinchBench measures how well LLM models perform as the brain of an OpenClaw agent. Results are collected on a public leaderboard at pinchbench.com. Prerequisites Python 3.10+ uv package manager OpenClaw instance (this agent) Quick Start Available Tasks (23) Task Category Description ------ ---------- ------------- task00sanity Basic Verify agent works task01calendar Productivity Calendar event creation task02stock Research Stock price lookup task03blog Writing Blog post creation task04weather Coding Weather script task05summary Analysis Document summarization task06events Research Conference research task07email Writing Email drafting task08memory Memory Context retrieval task09files Files File structure creation task10workflow Integration Multi-step API workflow task11clawdhub Skills ClawHub interaction task12skillsearch Skills Skill discovery task13imagegen Creative Image generation task14humanizer Writing Text humanization task15dailysummary Productivity Daily digest task16emailtriage Email Inbox triage task17emailsearch Email Email search task18marketresearch Research Market analysis task19spreadsheetsummary Analysis Spreadsheet analysis task20eli5pdfsummary Analysis PDF simplification task21openclawcomprehension Knowledge OpenClaw docs comprehension task22secondbrain Memory Knowledge management Command Line Options Option Description -------- ------------- --model Model identifier (e.g., anthropic/claude-sonnet-4) --suite all, automated-only, or comma-separated task IDs --output-dir Results directory (default: results/) --timeout-multiplier Scale task timeouts for slower models --runs Number of runs per task for averaging --no-upload Skip uploading to leaderboard --register Request new API token for submissions --upload FILE Upload previous results JSON Token Registration To submit results to the leaderboard: Results Results are saved as JSON in the output directory: Adding Custom Tasks Create a markdown file in tasks/ following TASKT …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aiskillstore/pinchbench