obsidian-wiki-ingest

name	obsidian-wiki-ingest
description	Ingest sources into the wiki.

Ingest Sources into the Wiki

Process sources into the persistent wiki at ~/Documents/wiki/.

Vault Layout

~/Documents/wiki/
├── raw/                   # Immutable source documents (user-placed, never modified by LLM)
│   ├── inbox/             # Staging — drop files/URLs here
│   ├── articles/          # Web articles, blog posts
│   ├── notes/             # Markdown, text files
│   ├── papers/            # PDFs
│   ├── videos/            # Video files
├── wiki/                  # LLM-maintained knowledge base
│   ├── concepts/          # Concept & topic pages
│   ├── entities/          # People, places, things
│   ├── sources/           # Source summaries
│   ├── synthesis/         # High-level overviews
│   └── analysis/          # Answers, comparisons, explorations
└── SCHEMA.md

Source Skepticism

Tier	Examples	Stance
High	Peer-reviewed papers, official docs	Trust unless contradicted
Medium	Known-author blogs, conference talks	Verify surprising claims
Low	Reddit/HN/forum posts, anonymous sources	Demand evidence

Flag unbacked strong claims: > ⚠️ **Unverified claim** from [source-type]: description — no evidence provided.

Separate observation from interpretation. Check for confounding factors. Never silently drop claims.

Inbox Ingest Flow

Trigger: /skill:obsidian-wiki-ingest (processes raw/inbox/) or /skill:obsidian-wiki-ingest <path-or-url>.

1. Detect and classify inbox contents → move to appropriate raw/ subdirectory
2. Duplicate check — wiki_search for existing coverage, check wiki/sources/<slug>.md
3. Read the source — read for files, web_fetch for URLs, yt-dlp for YouTube transcripts
4. Evaluate credibility — apply source skepticism tiers
5. Discuss takeaways — surface flagged claims for user review
6. Create/update pages — source summary, entities, concepts, synthesis
7. Empty inbox when done

YouTube transcript extraction

TMPDIR=$(mktemp -d)
yt-dlp --write-auto-sub --sub-format vtt --convert-subs srt --skip-download \
  --output "$TMPDIR/{id}.%(ext)s" {url}
sed '/^[0-9]/d; /^$/d; /-->/d' "$TMPDIR/{id}.en.srt" \
  | sed 's/<[^>]*>//g' | uniq > "$TMPDIR/{slug}.md"
# ... process, then:
rm -rf "$TMPDIR"

Page Conventions

• [[wiki links]] for all cross-references
• Filenames: lowercase-with-dashes
• Each page: title (H1), summary line, content, related links
• Mark contradictions: > ⚠️ Contradicts [[page]]: description