What problem does it solve? Valuable technical insights discovered during coding sessions—bug fixes, design decisions, version-specific behaviors—are lost when conversations end. This Skill reviews a conversation session, surfaces knowledge-worthy findings, and files them as GitHub issues in the dev-knowledge repository so they can be reused later. ## Core Features & Use Cases - Session Review: Scans the current conversation or past session logs to identify problems solved, non-obvious behaviors, design trade-offs, and reusable procedures. - User-Controlled Selection: Uses AskUserQuestion (or numbered plain-text fallback) to let the user choose the target conversation and which candidates to file—never auto-files everything. - Batch Issue Filing: Files each selected candidate as one topic per issue in dio0550/dev-knowledge with knowledge:pr or knowledge:direct labels, delegating the filing mechanics to the knowledge-issue skill. - Use Case: After a long debugging session, ask to harvest knowledge from the conversation; the Skill lists candidates like "React hook stale closure fix" and "Rust lifetime workaround", you pick which to keep, and each becomes a labeled GitHub issue. ## Quick Start Ask the AI to harvest knowledge from this conversation and file the selected findings as issues in the dev-knowledge repository.
