wiki-ingest

name	wiki-ingest
description	Ingest a raw source document into the wiki. Reads the source, extracts key information, creates or updates wiki pages, maintains cross-references, and logs the operation. Use when the user adds a new source or says 'ingest this'.
allowed-tools	Read Write Edit Glob Grep

Wiki Ingest

Process a new source document and integrate its knowledge into the wiki.

Workflow

Step 1: Read the source

Read the full source document from knowledge/. Identify:

• Key entities (people, organizations, concepts, technologies)
• Main claims and findings
• Relationships between entities
• Data points and statistics
• Contradictions with or confirmations of existing knowledge

Step 2: Discuss with the user

Before writing, share a brief summary of key takeaways. Ask:

• What aspects to emphasize?
• Any entities or concepts to prioritize?
• How granular should the decomposition be?
• What is this vault's purpose? (research, personal brain, coding docs, mentor search)

Skip this step if the user has already given clear instructions or if the vault's CLAUDE.md already defines the project scope.

Step 3: Update the wiki

For each significant entity or concept found in the source:

1. Check if a wiki page exists — read memory/wiki/index.md
2. If page exists — read it, integrate new information, update the "Sources" section, update frontmatter updated date and source_count
3. If page doesn't exist — create a new page in memory/wiki/ with proper frontmatter, content, and source citations
4. Update cross-references — add [[wikilinks]] in both directions between related pages

A single source typically produces 5-25 wiki pages. This is normal and expected. The AI 2027 article, for example, produced 23 pages (6 people, 5 organizations, 1 AI systems page, multiple concepts, and analysis pages).

Step 4: Write a source summary page

Create memory/wiki/sources/<source-name>.md with:

• One-paragraph summary
• Key claims extracted
• Entities mentioned (with wikilinks to their pages)
• Tags and metadata
• Date ingested

Step 5: Update index, hot cache, and log

1. Update memory/wiki/index.md — add/update entries for all pages touched, organized by category (tools, techniques, concepts, sources, people, comparisons)
2. Update knowledge/index.yaml — ensure the source is cataloged with tags and priority
3. (Optional) Update memory/wiki/hot.md — if the vault has a hot cache, update it with a ~500-word summary of what was just ingested
4. Append to memory/log.md:

   ## [YYYY-MM-DD] ingest | Source Title
   - Pages created: [list]
   - Pages updated: [list]
   - Key findings: [1-2 sentences]
   

Example

User drops knowledge/ai-2027-article.md and says "ingest this".

1. Read the article, identify entities: AI governance, OpenAI, model spec, Eli, Thomas, Daniel, superhuman coder, compute scaling
2. Discuss key findings with user (unless scope is already clear)
3. Create/update: ai-governance.md, openai.md, model-spec.md, eli.md, thomas.md, compute-scaling.md, etc.
4. Create: sources/ai-2027.md
5. Update index.md with all new/updated pages
6. Update hot.md with recent ingest summary (if hot cache is enabled)
7. Append ingest entry to log.md

Processing Expectations

• Single article: ~5-15 minutes depending on complexity
• Batch of 36 YouTube transcripts: ~14 minutes
• The LLM decides the taxonomy — flat or structured, whatever fits the content