Skip to content

feat(skills): add youtube-research-ingestor (INGEST)#1

Merged
kicrazom merged 1 commit into
mainfrom
feat/youtube-research-ingestor
Jun 21, 2026
Merged

feat(skills): add youtube-research-ingestor (INGEST)#1
kicrazom merged 1 commit into
mainfrom
feat/youtube-research-ingestor

Conversation

@kicrazom

Copy link
Copy Markdown
Owner

What

Adds the youtube-research-ingestor skill (INGEST stage): turns a YouTube URL into a local, source-graded research artifact — metadata, transcript (official subtitles first, local Whisper fallback), structured summary, an extracted-claims table, references/entities, a verification-needed list, optional technical-artifact + frame notes, and an audit.

Design

  • Local-first (yt-dlp + ffmpeg + whisper); hard preflight dependency gate before any run.
  • Treats YouTube as a secondary source; never promotes an unsupported claim to fact; never fabricates references/DOIs/timestamps/quotations.
  • Frame extraction = content-change (scene-detect) by default, not fixed intervals (ffmpeg select='gt(scene,THRESH)'), with a tunable threshold knob.
  • Composes with the rest of the toolkit: epistemic-status (claim grading), field-note-from-url (provenance note), literature-search (primary-source verification). Honors profile.knowledge_base.path.
  • Modes: transcript | summary | claims | scientific-review | technical-review | teaching-notes.

Changes

  • skills/youtube-research-ingestor/SKILL.md (new, markdown-only skill)
  • README.md — skills table row (INGEST)
  • CHANGELOG.md[Unreleased] → Added

No engine/code/test changes (markdown skill); composes with existing engines via the named sibling skills.

🤖 Generated with Claude Code

YouTube URL → local, source-graded research artifact: metadata, transcript
(official subtitles first, local Whisper fallback), structured summary,
extracted-claims table, references/entities, verification-needed list,
optional technical-artifacts + frame notes, and an audit.

Local-first (yt-dlp + ffmpeg + whisper). Treats YouTube as a secondary
source, never promotes an unsupported claim to fact, never fabricates
references/DOIs/timestamps/quotations. Composes with epistemic-status
(claim grading), field-note-from-url (provenance note), and
literature-search (primary-source verification). Modes: transcript |
summary | claims | scientific-review | technical-review | teaching-notes.

README skills table + CHANGELOG Unreleased/Added updated.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@kicrazom
kicrazom merged commit 360233a into main Jun 21, 2026
8 checks passed
@kicrazom
kicrazom deleted the feat/youtube-research-ingestor branch June 21, 2026 21:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant