RLM sub-LLM for segment-topics MAP. Reads ONE transcript-chunk file and returns topics + clip-worthy candidates (+ openthreads/callbacks) for that window only. Sonnet by default — topic tiling is cheap but the candidate ("most clip-worthy standalone moment") judgment is the quality-determining call Haiku under-rates.
Run the shorts pipeline across long-lived tmux panes. /start for a fresh run, or /start to reuse an already-ingested video. Each phase shows its progress in the foreground; tmux attach -t to inspect any phase live.
An agentic, vision-driven final QA/repair pass — a Claude "director" WATCHES the finished short (a contact sheet of frames sampled across the whole clip + the transcript + the sidecar plans) and decides, in natural language, what is broken ANYWHERE (cold open, dead/rambling tail, a tonally-wrong b-roll match, a…
The pre-fanout LEARNING GATE. start.sh finishes SPAN 0 alone (through the normal edit/captions/completion chain) before any lane launches, then this skill records a shot-scraper video demo of a generated per-span REVIEW PAGE (the 9:16 player + the grade/tier + the retention signals with the failing one flagged), has a…
Per-clip upload-readiness grade (0-99) read off a FINISHED .mp4 plus its persisted sidecar plans — the first skill that inspects a delivered pixel-level artifact, the on-disk proxy for YouTube's VVSA swipe gate. Deterministic retention-proxy floor (frame1isface, letterboxbars, creditlitatopen, firstvisualchange…
AGENTS.md instructions for jperrello/C0BALT_CUT, covering project instructions for ai agents, beads issue tracker, quick reference, rules and session completion.
Claude Code instructions for jperrello/C0BALT_CUT, covering project instructions for ai agents, beads issue tracker, quick reference, rules and session completion.