Add MetriLLM skill: benchmark local LLM models

[TheBlueHouse75]

Summary

• Adds MetriLLM skill to the ClawHub registry
• MetriLLM is a CLI benchmark tool for local Ollama LLM models
• Measures performance (tokens/second, time to first token, memory usage) and quality (reasoning, math, coding, instruction following, structured output, multilingual)
• Computes a hardware fitness verdict: EXCELLENT / GOOD / MARGINAL / NOT RECOMMENDED

Skill details

Field	Value
Name	metrillm
Author	MetriLLM
License	MIT
Source	https://github.com/MetriLLM/metrillm
Required tools	Bash, Read
Dependencies	Ollama

What users can do with this skill

• Run full benchmarks (npx metrillm bench --model <name>) measuring both performance and quality
• Run quick performance-only benchmarks (--perf-only)
• View previous results stored in ~/.metrillm/results/
• Share results to the public leaderboard (--share)

Test plan

• SKILL.md follows the ClawHub skill format with valid YAML frontmatter
• Skill directory follows skills/<owner>/<slug>/SKILL.md convention
• Content fetched from upstream repo: https://raw.githubusercontent.com/MetriLLM/metrillm/main/SKILL.md

🤖 Generated with Claude Code

[openclaw-barnacle]

Thanks for the pull request! This repository is read-only and is automatically synced from https://clawhub.ai, so we can’t accept changes here. Please make updates on the website instead.