Review

Skill avanrossum/claude-local-llm-skills/skills/review

Run a first-pass code review using the local LLM — catches obvious issues without spending Claude tokensFrom its SKILL.md

Install

npx -y skills add avanrossum/claude-local-llm-skills --skill review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.4 KB, as published. Nobody here has run it

/review — Code Review via Local LLM

You are running a first-pass code review using the local/remote LLM. The local model catches obvious issues (bugs, style, missing error handling); the user can escalate to you for deeper architectural review.

Arguments: $ARGUMENTS

Step 1 — Determine What to Review

Parse the argument:

A file path: review that file
staged or diff: run git diff --cached to get staged changes, review those
last or HEAD: run git diff HEAD~1 to review the last commit
No argument: default to staged changes

Step 2 — Send to Local LLM

Read the file content or diff, then call the local LLM:

LLM_API_TOKEN=$(security find-generic-password -a "$USER" -s "lm-studio-token" -w 2>/dev/null)
curl -s "${LLM_BASE_URL:-https://lm.mipyip.com}/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $LLM_API_TOKEN" \
  -d '{
    "model": "'"${LLM_MODEL:-qwen/qwen3.5-35b-a3b}"'",
    "messages": [
      {"role": "system", "content": "You are a senior code reviewer. Review the following code and identify:\n1. Bugs or logic errors\n2. Missing error handling\n3. Security concerns\n4. Style issues or anti-patterns\n5. Suggestions for improvement\n\nBe concise. Only flag real issues, not nitpicks. Format as a numbered list."},
      {"role": "user", "content": "{content}"}
    ],
    "stream": false,
    "temperature": 0.3
  }'

Environment variables:

LLM_BASE_URL — override endpoint (default: https://lm.mipyip.com)
LLM_MODEL — override model (default: qwen/qwen3.5-35b-a3b)
Bearer token from macOS Keychain (lm-studio-token) or LLM_API_TOKEN env var

Step 3 — Present Results

Relay the local model's findings to the user. Add a brief note at the end:

This review was performed by the local LLM ({model name}). For deeper architectural analysis, ask me directly.

If the local model found no issues, say so — don't invent problems.

Error Handling

If the LLM is unreachable, offer to review the code yourself
If the diff/file is too large, split it and send in chunks, or summarize which files changed and offer to review specific ones