111 lines
4.7 KiB
Markdown
111 lines
4.7 KiB
Markdown
|
---
|
||
|
|
pageType: source
|
||
|
|
id: source.personal-knowledge-companion
|
||
|
|
title: personal-knowledge-companion
|
||
|
|
sourceType: local-file
|
||
|
|
sourcePath: /home/topher/.openclaw/workspace-crash-bot/memory/personal-knowledge-companion.md
|
||
|
|
ingestedAt: 2026-05-03T01:07:44.335Z
|
||
|
|
updatedAt: 2026-05-03T01:07:44.335Z
|
||
|
|
status: active
|
||
|
|
growth: seed
|
||
|
|
---
|
||
|
|
|
||
|
|
# personal-knowledge-companion
|
||
|
|
|
||
|
|
## Source
|
||
|
|
- Type: `local-file`
|
||
|
|
- Path: `/home/topher/.openclaw/workspace-crash-bot/memory/personal-knowledge-companion.md`
|
||
|
|
- Bytes: 4041
|
||
|
|
- Updated: 2026-05-03T01:07:44.335Z
|
||
|
|
|
||
|
|
## Content
|
||
|
|
```text
|
||
|
|
# Personal Knowledge Companion / Life View Dashboard
|
||
|
|
|
||
|
|
## What It Is
|
||
|
|
A persistent knowledge layer across all agents that:
|
||
|
|
1. Maps what -topher knows, what's being learned, and where the gaps are
|
||
|
|
2. **Proactively tells him about those gaps** before he realizes he needed them
|
||
|
|
3. Groups knowledge by life domain (School/Work/Play)
|
||
|
|
|
||
|
|
Not a passive archive. An active learning companion that watches activity, builds a model of knowledge edges, and surfaces learning opportunities without being asked.
|
||
|
|
|
||
|
|
## Core Components
|
||
|
|
|
||
|
|
### 1. memory-wiki (OpenClaw Plugin)
|
||
|
|
Built into OpenClaw. Plain `.md` files, Obsidian-compatible.
|
||
|
|
- **Vault layout:** `entities/`, `concepts/`, `syntheses/`, `sources/`, `reports/`
|
||
|
|
- **Structured claims** with evidence, confidence, contradictions
|
||
|
|
- **Per-agent vault mode** — separate wiki per agent, isolated knowledge domains
|
||
|
|
- **Bridge mode** — pulls from active memory plugin
|
||
|
|
- Built-in dashboards: open questions, contradictions, low-confidence claims, stale pages, relationship graphs
|
||
|
|
- Tools: `wiki_search`, `wiki_get`, `wiki_apply`
|
||
|
|
- Docs: `https://docs.openclaw.ai/plugins/memory-wiki`
|
||
|
|
|
||
|
|
**Gap detection is the key feature:** The wiki tracks topics covered, topics stale, topics never explored. Reports surface automatically.
|
||
|
|
|
||
|
|
### 2. Domain Separation
|
||
|
|
- 🏫 **School** — 2890 bot (robotics team)
|
||
|
|
- 🏭 **Work** — PSB bots (brewery ops)
|
||
|
|
- 🎮 **Play** — crash-bot / HHS-Hackers crew
|
||
|
|
|
||
|
|
Each domain has its own vault that can cross-reference the others.
|
||
|
|
|
||
|
|
### 4. Professor Agent (The Teacher) — Scales to Anyone
|
||
|
|
Same knowledge graph, different personal layers per person.
|
||
|
|
- Person entities: confidence scores, scale preferences, learning style, delivery channel
|
||
|
|
- When Kyle wants CEH prep → professor creates entity/kyle.md, assesses current knowledge, builds a learning path
|
||
|
|
- When Bruno needs PathPlanner → professor creates entity/bruno.md, starts from his mechanical strength, fills gaps
|
||
|
|
- Each person's learning path compounds into the shared knowledge graph for the next person
|
||
|
|
- Access control maps to existing Discord structure (HHS-Hackers, 2890, DMs)
|
||
|
|
|
||
|
|
### 4. Playful Visual Layer (Inspired by Claw Empire)
|
||
|
|
- Claw Empire: pixel-art office simulator, agents as employees in a virtual company
|
||
|
|
- Adapted for personal life context instead of a coding shop
|
||
|
|
- Could visualize domains, agent status, knowledge flow
|
||
|
|
- Not full Claw Empire — just the playfulness and visual mapping
|
||
|
|
|
||
|
|
## Related Concepts
|
||
|
|
- **Karpathy's LLM Wiki pattern** — raw/ notes → AI synthesizes → wiki pages with auto-links
|
||
|
|
- **AgentWiki** (agentwiki.org) — shared knowledge base for AI agents, JSON-RPC API
|
||
|
|
- **Obsidian** — same plain-Markdown vault format, compatible
|
||
|
|
- **SamurAIGPT/llm-wiki-agent** — personal knowledge base that builds itself
|
||
|
|
- **kytmanov/obsidian-llm-wiki-local** — 100% local, Ollama-powered
|
||
|
|
|
||
|
|
## Claw Empire Reference
|
||
|
|
- GitHub: `GreenSheep01201/claw-empire`
|
||
|
|
- "Command Your AI Agent Empire from the CEO Desk"
|
||
|
|
- Pixel-art office metaphor, git worktrees, agent meetings and deliverables
|
||
|
|
- Not a coding shop fit — over-engineered for -topher's use case
|
||
|
|
- Inspiration for visual/playful layer, not the architecture
|
||
|
|
|
||
|
|
## Life View Project File
|
||
|
|
See: `projects/life-view-dashboard.md`
|
||
|
|
|
||
|
|
## The Thread
|
||
|
|
Needs its own Discord channel: `#personal-knowledge-companion` or `#life-view`
|
||
|
|
|
||
|
|
## Status
|
||
|
|
Concept stage. Not built yet. Needs:
|
||
|
|
- [x] Discord channel (#personal-knowledge-companion)
|
||
|
|
- [ ] memory-wiki configuration
|
||
|
|
- [ ] Professor agent (replaces librarian)
|
||
|
|
- [ ] Gap detection logic (driven by wiki claims + confidence scores)
|
||
|
|
- [ ] Learning path delivery (Discord training threads)
|
||
|
|
- [ ] Knowledge graph visualization (Obsidian)
|
||
|
|
|
||
|
|
## Why It Matters
|
||
|
|
Current agents are mostly command-and-reply. They don't do research, self-direct, or maintain persistent context between sessions. They feel like fancy autocomplete, not assistants.
|
||
|
|
|
||
|
|
This gives them continuity, memory, and proactivity — and gives -topher a way to see his whole digital life at a glance and understand where his knowledge gaps are.
|
||
|
|
```
|
||
|
|
|
||
|
|
## Notes
|
||
|
|
<!-- openclaw:human:start -->
|
||
|
|
<!-- openclaw:human:end -->
|
||
|
|
|
||
|
|
## Related
|
||
|
|
<!-- openclaw:wiki:related:start -->
|
||
|
|
- No related pages yet.
|
||
|
|
<!-- openclaw:wiki:related:end -->
|