---
namespace: work
tags:
  - wiki
  - schema
  - meta
created: '2026-05-13'
updated: '2026-05-13'
---
# Wiki Schema — Personal OS Knowledge Base

This file defines the domain, structure rules, and curation guidelines for the `wiki/` folder.
Used by the daily Eagle wiki-curation cron job.

## Domain

This wiki covers Alex's personal operating system — the tools, workflows, decisions, and architecture
that run his work and personal life. It is NOT a general knowledge base — everything here must be
relevant to Alex's specific setup.

## Existing Pages

| File | Topic | Namespace |
|------|-------|-----------|
| `personal-os-architecture.md` | Full system architecture (Eagle, Taiga, Kraken, pipelines) | work |
| `personal-os-schema.md` | Postgres DB schema (personal_os) | work |
| `personal-os-purpose.md` | Why the system exists, goals | work |
| `personal-os-agent-rules.md` | Agent behavior rules and constraints | work |
| `personal-os-self-modification.md` | How the system evolves and self-improves | work |
| `personal-os-sync-pipeline.md` | Asana sync, data pipeline internals | work |
| `personal-os-catchup-plan-2026-04-27.md` | Catchup plan snapshot (April 2026) | work |
| `personal-os-state-2026-04-27.md` | System state snapshot (April 2026) | work |
| `ddg-asana-workflow.md` | DuckDuckGo + Asana workflow | work |
| `obsidian-mcp-wrapper.md` | obsidian MCP / mcpvault setup notes | work |
| `vault-filling-guide.md` | How to populate the vault | work |
| `wiki-ingest-process.md` | Wiki ingest pipeline description | work |
| `tech/media-pipeline-pitfalls.md` | Swift media-pipeline pitfalls (§9.x, transliteration, Docker run) | work |
| `tech/vault-namespace.md` | Vault namespace rules: work/personal/family separation | work |
| `tech/kraken-network.md` | Kraken network topology, WireGuard, media volume mounts | work |
| `tech/htpc-steam-emulators.md` | HTPC Steam/emulator rules, gamepad config, save paths | personal |
| `concepts/agent-memory-architecture.md` | Three-level memory model for LLM agents; Hermes implementation | work |
| `concepts/vault-strategy.md` | Three vault tasks: enrichment, wiki accumulation, proactive research | work |
| `concepts/kraken-media-stack.md` | Full Kraken media server: media-pipeline + *arr + Jellyfin; pitfalls synthesis | work |
| `entities/library-app.md` | Personal book library UI on TrueNAS (687k books, deployed 2026-05-16) | personal |

## Curation Rules

1. **One topic = one file.** Never merge unrelated topics.
2. **Wiki pages are synthesized, not copied.** Summarize, don't paste raw content.
3. **Update existing pages** rather than creating duplicates.
4. **Cross-reference** related pages with `[[filename]]` links.
5. **Frontmatter required** on every page:
   ```yaml
   ---
   namespace: work | personal | family
   tags: [tag1, tag2]
   last_synced: YYYY-MM-DD
   confidence: 0.0–1.0
   ---
   ```
6. **Never delete** pages — mark stale content with `> ⚠️ Stale as of YYYY-MM-DD`.
7. **Confidence** < 0.5 = speculation, add disclaimer. > 0.8 = verified fact.

## Sections

| Subfolder | What goes here |
|-----------|---------------|
| *(root wiki/)* | personal-os infrastructure, architecture, agent rules |
| `concepts/` | abstract topics, tech concepts, methodologies |
| `entities/` | tools, products, services, people |
| `tech/` | development topics relevant to Alex's stack — pitfalls, debug notes, infra patterns |
| `ideas/` | ideas, hypotheses, open questions |
| `queries/` | filed answers to valuable one-off questions |
| `comparisons/` | side-by-side analyses |
| `raw/inbox/` | drop zone — files here get processed by nightly cron |
| `raw/articles/` | extracted web articles (immutable after write) |
| `raw/papers/` | PDFs and papers (immutable after write) |
| `research-queue.md` | proactive research queue — topics to investigate (Saturday cron) |

## Inbox Protocol

Files in `raw/inbox/` are processed nightly by the curation cron:
- `research-[slug]-YYYYMMDD.md` — research request
- `idea-YYYYMMDD-HHMMSS.md` — idea to preserve
- `url-YYYYMMDD-HHMMSS.md` — URL to extract and synthesize
- `note-YYYYMMDD-HHMMSS.md` — plain note or decision
After processing, move to `raw/inbox/processed/`.

## What to Curate

- Architecture changes (new services, infra decisions)
- New workflows or agent capabilities
- Lessons learned from system failures or fixes
- Updated tool versions, API changes
- New integrations (Tavily, mcpvault, etc.)
- Anything dropped into `raw/inbox/` by Alex or Kraken

## What NOT to Add

- Raw task lists (those live in Asana)
- Personal/family content (those live in `personal/` and `family/`)
- General knowledge that isn't Alex-specific (exception: `tech/` and `ideas/` — those are fine)