extract-learnings

To extract coding guidelines from PR feedback, run the extract-learnings workflow on a target PR. It parses review comments, filters out bot messages, and saves actionable team conventions directly to your project's general_coding_instructions.md file.

Yes, the extract-learnings workflow supports both GitHub and Azure DevOps pull requests. It parses PR feedback from either platform, filters out system and bot comments, and extracts structured engineering learnings with source references.

Documenting team conventions from code review comments is best handled by categorizing extracted learnings into existing sections like Code Style or Error Handling. This skill automatically updates your .github/general_coding_instructions.md file to codify these patterns.

You can update general_coding_instructions.md by running the extract-learnings workflow after a PR review concludes. The skill automatically categorizes actionable takeaways and applies them to existing or newly created sections in the project documentation.

Yes, you can filter out bot and system messages when parsing PR feedback. The extract-learnings workflow automatically removes these automated comments to ensure only high-quality, human-generated engineering learnings are captured and saved.

You should run the workflow to capture engineering learnings when PR reviews conclude. Running it at this stage ensures all feedback on naming, formatting, and error handling is parsed and codified into the project's general_coding_instructions.md file.