andrewleech

Use this skill for general MicroPython device interaction via mpremote, including connecting, running code, checking device state, and managing the device. Triggers on "connect to micropython", "run code on device", "check device state", "mpremote", "micropython device", "repl", "device version", "device reset".

Use this skill when copying files to or from a MicroPython device using mpremote. Covers file copy syntax, directory operations, and best practices for device filesystem management. Triggers on "copy file to device", "upload to device", "download from device", "mpremote fs", "device filesystem", "copy firmware files".

Use this skill when maintaining a persistent interactive connection to a MicroPython device for sending commands and capturing output. This is the PREFERRED method for device interaction when multiple commands need to be sent, when monitoring device output over time, or when the device runs an asyncio event loop with aiorepl. Triggers on "send commands to device", "monitor device output", "interactive session", "persistent connection", "stress test device", "capture serial output".

Review MicroPython code changes using domain-focused agents. Invoke when user mentions reviewing code, wants feedback on MicroPython PRs/commits/diffs, or asks for code review.

This skill MUST be used when drafting GitHub Pull Request or GitLab Merge Request titles and descriptions. This includes when the user asks to "create a PR", "open a PR", "raise a MR", "write a PR description", "draft a merge request", "submit a PR", "make a merge request", or when an agent workflow reaches a PR/MR creation step. This skill should also be used when reviewing or rewriting an existing PR/MR description.

Run the MicroPython CI pipeline locally using Docker. This skill should be used when working in a MicroPython repository and the user wants to run CI checks, build firmware, or test changes locally before pushing. Triggers on requests like "run CI", "test my changes", "build stm32", "run all checks", "run the linters", or any mention of local CI testing for MicroPython.

Systematic workflow for maintaining a backlog of open MicroPython PRs. Use when the user asks about PR maintenance, PR backlog processing, rebasing PRs, working through open PRs, or updating stale PRs. Handles reconnaissance, reviewer feedback discussion, rebase in worktrees via subprocess, local CI validation, and force-push.

This skill should be used when the user wants to list, sort, or triage open MicroPython pull requests. Invoke when user mentions open PRs, PR triage, PR status, checking feedback on PRs, or wants to see which PRs need attention. Shows branch names, external feedback status, and local worktree/branch availability.

Use this skill when the user wants to review MicroPython code changes using historical review patterns. Invoke when user mentions reviewing code, finding review examples, or wants feedback on MicroPython PRs/commits/diffs. Provides semantic search across ~19.5K categorized review comments.

Use this skill when the user wants to review MicroPython code changes using dpgeorge's review patterns. Invoke when user mentions reviewing code, finding review examples, or wants feedback on MicroPython PRs/commits/diffs in dpgeorge's style. Provides semantic search across 18,614 categorized review comments.

Run structured model optimization experiments for Qwen3-ASR ONNX. Use when testing quantization variants, encoder formats, weight sharing, or other model changes that need WER/RTF measurement. Triggered by requests to "run an experiment", "test a variant", "try quantizing", "measure WER", or any model optimization trial in the qwen3-asr-onnx repo.

This skill should be used when the user wants to find duplicate or related MicroPython issues/PRs, triage an issue for similarity to existing items, detect spam or off-topic issues, or check if a new issue has already been reported. Invoke when user mentions finding duplicates, triaging issues, checking for related PRs, or detecting spam.

Launch a serve-review web UI code review gate. This skill should be used when the user asks to "serve a review", "launch a serve-review", "local code review", "review before pushing", or any variant of starting a web-based local code review session.

Search past conversation history and answer questions about previous project discussions. Use when asked about "how did we...", "what was our approach to...", or any question requiring context from past conversations. On first use, check installation status.