Lore DB | Docs

Audit & Verify Content

Review documents for staleness, consolidate duplicates, and verify important documentation. This is the “weekly review” workflow for knowledge champions.

When to use this

  • Weekly: Review the Pending Insights library (15 min)
  • Monthly: Audit a team library for stale content
  • Quarterly: Comprehensive review of Core Playbooks

The weekly review (Pending Insights)

Copy-paste prompt
Time for the weekly knowledge review. Please help me review
the Pending Insights library.

1. List all documents in our Pending Insights library using
 search_doc with the library_id
2. For each document, read it and categorize it:
 - PROMOTE: Verified, useful → tell me which Team or Core
   library it should move to
 - MERGE: Same topic as existing doc → identify the target
   document to merge into
 - ARCHIVE: Too narrow, outdated, or incorrect → recommend
   deletion
 - KEEP: Needs more time or information before deciding
3. Summarize your recommendations in a table:
 | Document | Action | Target | Reason |

I'll approve each action before you execute it.

Monthly library audit

Copy-paste prompt
Please audit the [LIBRARY NAME] library for content quality.

1. List all documents in the library
2. For each document, check:
 - Is the title descriptive and searchable?
 - Is the content well-structured with headings?
 - Does it cover one focused topic?
 - Is there any obviously outdated information?
3. Flag any documents that should be:
 - Renamed (poor title)
 - Split (covers multiple unrelated topics)
 - Merged (duplicate or overlapping content)
 - Verified (content confirmed still accurate)
 - Deleted (no longer relevant)
4. After review, verify any documents I approve using verify_doc

Document hygiene rules

RuleWhy
One topic per documentSemantic search works best with focused chunks
Descriptive titlesTitles are the strongest search signal
No orphan documentsEvery doc belongs to at least one shared library
Consolidate, don’t accumulateFive small notes < one structured document
Verify quarterlyUnverified documents get deprioritized in search

Naming convention

Use the pattern: [System/Service] — [Specific Topic]

  • Payment Service — Stripe Webhook Retry Configuration
  • Infrastructure — Kubernetes Pod Autoscaling Thresholds
  • Onboarding — New Engineer First Week Checklist
  • Incident — 2026-03-15 Database Connection Pool Exhaustion

The biggest risk isn’t too little documentation — it’s too much low-quality documentation that makes search results noisy. The weekly review is the most important habit to form.