Process work sessions into the claude-code-plugins knowledge graph — preserves the session transcript, writes or links its chronicle entry as the session's synthesis, turns the transcript into observations, updates the patterns and decisions they support, folds in named plugins and skills, and proposes promotions to the personal commons. Also resolves-and-preserves an on-demand Claude Code docs URL when one comes up. Use when the user says "process this session", "run /process", names a session transcript or UUID, names a chronicle entry whose session should be processed, asks what's in the processing queue, or points at a Claude Code release-notes / docs page to capture.
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
The skill workflow processes work sessions by ingesting session transcripts from local files (`*.jsonl`) and optionally fetching Claude Code documentation pages via WebFetch URLs, exposing the agent to outsider-authored free text from session contents.
24e103e
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.