What problem does it solve? Technical work with AI agents often loses context between sessions and picks the wrong specialized skill for the job. This Skill classifies incoming technical requests, selects the minimal set of specialized skills needed, and persists decisions, artifacts, and preferences to a shared GitHub memory repository so context survives across sessions. ## Core Features & Use Cases - Task Routing: Classifies requests into programming, design, full stack, databases, security, or game development domains and maps them to a curated routing matrix of external skills. - Persistent Memory Protocol: Enforces a mandatory session-start and session-end sync with the authorized GitHub repository, writing structured Markdown entries into categorized monthly files. - Safe Git Automation: Ships scripts that validate the remote origin, restrict writes to the Manus/ directory, redact credentials, and optionally commit and push memory updates. - Use Case: A developer asks for help debugging a React app. The Skill loads prior project context from the memory repository, routes to systematic-debugging and react-best-practices, verifies the fix, then records the decision and pending items back to the repository. ## Quick Start Classify my request to fix a failing API endpoint, load any relevant project memory from the repository, pick the right specialized skill, and record the outcome when finished.
