What problem does it solve? Community maintainers need to respond to GitHub issues and discussions consistently and safely, but fully automated posting risks wrong or off-brand replies. This Skill scans open community items, drafts templated responses, and enforces a human approval gate before anything is posted. ## Core Features & Use Cases - Community Scanning: Finds unanswered open issues and discussions from the last 7 days using GitHub MCP tools or the gh CLI, excluding internal or wontfix items. - Classification and Templated Drafting: Classifies each thread into nine response types (welcome, troubleshooting, redirect, and more) and drafts from matching templates with confidence levels and auto-escalation rules. - Human Review Gate and Audit Trail: Presents drafts in a review table, waits for explicit approval commands, supports a freeze safe word, and logs every action to an audit directory. - Use Case: A maintainer says "PAO, check community" and receives a table of drafted replies to recent issues and discussions, approves two of them, and posts them via prepared gh CLI commands with full audit logging. ## Quick Start Ask the agent to check the community by saying "PAO, check community" to scan recent GitHub issues and discussions and receive drafted responses for review.