Extract a comprehensive operator-mention analysis from each new video transcript. Reads the SRTs in .help/.tmp/video-transcripts/ (produced by video_to_srt.py), and for each one without an analysis yet, writes references/video-analysis/<id>.md — every meaningful operator discussion with its timestamp, depth, and a one-line note. Feeds the editor's deep-link index; never touches the wiki, the YouTube descriptions, or git. Use when the user wants to analyze transcribed videos, build the operator mention index, or invokes /analyze-videos.
Draft the embedded help text for ui: topics whose doc is still empty. For each topic it gathers the transcript passages where the maintainer explains it (via the ui: deep-link index), the implementing class, and any existing .help/ page, then distills a short user-facing doc into .help/references/topics/ui-topics.md. Use after the videos are transcribed and analyzed (stage 2), when the user wants to write or fill in the ui topic docs.
Unattended batch over the "In progress" column of the TiXL road-map GitHub board. Implements straightforward tickets as individual git stashes and writes plans for the rest as files in .agentic/Plans/ — no questions, no commits, no GitHub writes. Use when the user wants to work through their In-progress tickets, process the board backlog, or invokes /process-tickets.
Walk through open Sentry issues for tooll3, latest first, proposing a fix and committing each one separately with user review between. Use when the user wants to triage Sentry, fix Sentry issues, work through the Sentry backlog, or invokes /sentry-fix.