Reading the Corpus Health Report

The corpus health report gives you the overhead map of the knowledge base — which shelves are full, which are empty, where pathways exist and where they don't. Run it to understand the current state and decide what to do next.

Running the Report

From the repository root:

pnpm knowledge:health

No arguments needed. The report scans all documents in the knowledge/ directory (sources, commentary, references, guides, collections) and prints a structured analysis to the terminal.

Report Sections

Overview

── Overview ────────────────────────────────────────────
   Documents:       11
   Domains active:  3 / 16
   Cross-refs:      5
   Graph density:   4.5%

• Documents: Total number of published documents across all roles (sources, guides, collections, etc.).
• Domains active: How many of the 16 available domains have at least one document. The full domain list: buddhism, stoicism, sufism, taoism, vedic, indigenous, philosophy, ecology, science, psychology, literature, art, engineering, ai, opencosmos, cross.
• Cross-refs: Total count of all related_docs references across the corpus.
• Graph density: Cross-refs as a percentage of maximum possible connections. A fully connected graph of N documents would have N*(N-1) directed edges. Early in the corpus's life, density will be low — that's expected.

Domain Coverage

── Domain Coverage ─────────────────────────────────────
   opencosmos     ████████████████████ 5
   cross          ████████████ 3
   buddhism       ████ 1
   stoicism        (empty)
   sufism          (empty)
   ...

A visual bar chart showing how many documents exist in each domain. Empty domains are flagged. Use this to identify gaps — if a foundation collection references stoic texts but the stoicism domain is empty, that's a priority.

Role Coverage

── Role Coverage ───────────────────────────────────────
   source         3
   commentary     0
⚠️ reference      0
   guide          5
   collection     4

How many documents exist in each role category. A warning icon appears for roles with zero documents. A healthy corpus has sources as the foundation, commentary building on sources, and references for quick lookup.

Foundation Collection Progress

── Foundation Collection Progress ──────────────────────
   sol foundations
   ░░░░░░░░░░░░░░░░░░░░ 0/12 (0%)
   socrates foundations
   ░░░░░░░░░░░░░░░░░░░░ 0/10 (0%)
   optimus foundations
   ░░░░░░░░░░░░░░░░░░░░ 0/8 (0%)
   cosmo foundations
   ░░░░░░░░░░░░░░░░░░░░ 0/6 (0%)

Each foundation collection has placeholder entries (- [ ] Text Title) for texts that need to be imported. This section shows progress bars — how many placeholders have been filled (converted to - [x] links by the publication CLI) versus how many remain.

When a source text is published whose title matches a placeholder, the CLI automatically checks the box and adds a link. This section tracks that progress across all four collections.

Cross-Reference Integrity

── Cross-Reference Integrity ───────────────────────────
   ✅ All cross-references resolve correctly.

or:

   ⚠️  2 broken references:
      sources/buddhism-the-dhammapada.md → commentary/deleted-doc.md (not found)

Validates that every related_docs path in every document's frontmatter points to a file that actually exists. Broken references happen when documents are renamed, moved, or deleted without updating the documents that link to them.

Islands (Zero Connections)

── Islands (Zero Connections) ──────────────────────────
   ⚠️  3 islands:
      The Dhammapada (sources/buddhism-the-dhammapada.md)
      The Prophet (sources/cross-the-prophet.md)

A document is an "island" if no other document references it AND it references no other documents. Islands are disconnected from the knowledge graph — they exist but aren't woven into the web of relationships.

To fix an island, add it to relevant documents' related_docs lists, or add related_docs to the island document's own frontmatter.

Import Priority

── Import Priority (Top Texts to Add Next) ────────────
   1. Meditations (Marcus Aurelius)
      Referenced in: sol foundations, socrates foundations | Score: 5
   2. Tao Te Ching
      Referenced in: sol foundations, cosmo foundations | Score: 4
   3. Rumi — Selected Poems
      Referenced in: sol foundations | Score: 3

The top texts to import next, scored by:

• Collection mentions (+1 per foundation collection that references the text)
• New domain bonus (+2 if importing the text would create a new active domain)

Texts that appear in multiple foundation collections score highest — they're cross-domain bridge texts that serve multiple voices in the AI Triad.

When to Run It

• After publishing: See the impact of your new addition on domain coverage and graph density.
• Planning what to import next: The import priority section tells you which texts have the highest leverage.
• Periodic check-up: Run it weekly or before a batch of imports to understand the current landscape.
• After refactoring: If you've renamed or reorganized documents, check cross-reference integrity.

Acting on the Report