April 19, 2026 · Git · Filesystem
TeamDay — Knowledge Base Audit
3945 docs · 23.1 MB of institutional memory, classified, aged, and ranked for review.
Produced by Priya — HR Manager — Knowledge Base & People Ops
Fresh (≤7d)
207
updated this week
Stale (>60d)
462
needs review
What's in the knowledge base
Scanned 3945 markdown docs across the TeamDay monorepo. Strategy notes, runbooks, CLAUDE.md files, skill definitions, subagent specs. If the team's collective memory lives anywhere, it's here.
Docs by type
Coverage by document type
| Type | Count | Total size | Avg age (days) |
|---|
| Other | 3,422 | 20432 KB | 53 |
| Docs | 407 | 2603 KB | 50 |
| Skill | 54 | 280 KB | 41 |
| README | 49 | 301 KB | 114 |
| CLAUDE | 11 | 28 KB | 75 |
| Subagent | 2 | 8 KB | 20 |
Stale — not touched in 60+ days (top 12 by age)
| Path | Age (days) | Size |
|---|
| .vscode/FORMATTING_GUIDE.md | 160 | 3.7 KB |
| codex-agents/README.md | 160 | 2.0 KB |
| codex-agents/cloud-cost-optimizer.agent.md | 160 | 2.9 KB |
| codex-agents/competitor-intelligence-analyst.agent.md | 160 | 2.7 KB |
| codex-agents/documentation-knowledge-steward.agent.md | 160 | 2.3 KB |
| codex-agents/performance-optimization-analyst.agent.md | 160 | 2.7 KB |
| codex-agents/product-strategy-director.agent.md | 160 | 2.9 KB |
| codex-agents/security-remediation-engineer.agent.md | 160 | 2.4 KB |
| codex-agents/security-risk-analyst.agent.md | 160 | 2.8 KB |
| codex-agents/workflow-automation-architect.agent.md | 160 | 2.7 KB |
| docs/archive/AGENT_CEO.md | 160 | 9.4 KB |
| docs/archive/FILE_UPLOAD_SECURITY.md | 160 | 5.2 KB |
Biggest knowledge files
| Path | Size | Lines | Headings | Age |
|---|
| .claude/skills/prompt-tracker/output/latest-upstream.md | 65.2 KB | 918 | 41 | 3d |
| docs/archive/technical/implementations/hooks-system.md | 64.2 KB | 2,335 | 55 | 160d |
| .claude/skills/prompt-tracker/output/proposed-update.md | 63.7 KB | 922 | 40 | 3d |
| docs/tasks/retrospective-ux-design.md | 43.9 KB | 1,271 | 45 | 82d |
| docs/strategy/character-microapps-primer.md | 43.0 KB | 952 | 69 | 65d |
| docs/archive/technical/implementations/HOOKS_ARCHITECTURE.md | 36.3 KB | 650 | 20 | 160d |
| docs/202602/ai-team-architecture.md | 34.8 KB | 706 | 47 | 67d |
| packages/marketing/content/ja/blog/ai-sales-agent-deployment-guide.md | 34.2 KB | 238 | 9 | 25d |
What this tells Priya
Freshness pulse
207 docs updated in the last 7 days — the team is actively writing. 462 docs haven't changed in 60+ days — mark for review.
Stub risk
3 docs are under 200 bytes — likely placeholders or orphans. Audit and either expand or delete to stop polluting search.
Coverage gaps
Review the type histogram: if runbooks or onboarding guides are thin, the next hire's first week will suffer. Fill gaps proactively.
Living memory
CLAUDE.md files are the team's agent-facing memory. Count them per package — every long-lived package should have one.
Recommendations
- Owner every doc in the "Stale" list this sprint — either refresh it, archive it, or mark it canonical with a "last verified" line.
- 3 stubs exist — Priya should run a monthly "stub amnesty" pass: expand to 300+ words, demote to a comment, or delete.
- Add a "knowledge freshness" reminder to the Friday retro. Team sees their own docs aging; self-corrects.
- When a new hire joins, auto-generate a reading list from the top-20 newest-and-largest docs filtered to their area — saves a week of onboarding.
- Rerun this audit monthly; diff counts of stale/fresh to measure if the org is losing or gaining institutional memory.