vault-sync: 2026-09-13 12:10 (jimmy)
This commit is contained in:
jimmy
@@ -1,11 +1,12 @@
|
||||
# Vault Updates Log
|
||||
|
||||
**Generated:** 2026-09-13 12:07 PST
|
||||
**Generated:** 2026-09-13 12:10 PST
|
||||
|
||||
---
|
||||
|
||||
## Last 24 Hours
|
||||
|
||||
aa52eec vault-sync: 2026-09-13 12:07 (jimmy)
|
||||
e918f23 vault-sync: 2026-09-13 12:01 (jimmy)
|
||||
32ee89f vault: drop stray git-credential-manager sentinel artifact
|
||||
73d0bb9 vault-sync: 2026-09-13 12:00 (jimmy)
|
||||
@@ -17,6 +18,7 @@ fcf672f vault-sync: 2026-09-13 11:53
|
||||
|
||||
## Last 7 Days
|
||||
|
||||
aa52eec vault-sync: 2026-09-13 12:07 (jimmy)
|
||||
e918f23 vault-sync: 2026-09-13 12:01 (jimmy)
|
||||
32ee89f vault: drop stray git-credential-manager sentinel artifact
|
||||
73d0bb9 vault-sync: 2026-09-13 12:00 (jimmy)
|
||||
@@ -33,6 +35,7 @@ fcf672f vault-sync: 2026-09-13 11:53
|
||||
01_Wiki/ai-bottleneck-thesis-tony-simons.md
|
||||
01_Wiki/hermes-masterclass-tonbi.md
|
||||
01_Wiki/inference-benchmarks-local.md
|
||||
01_Wiki/omniroute-routing-failover.md
|
||||
02_agents/jimmy/README.md
|
||||
02_agents/leonard/lessons.md
|
||||
02_agents/leonard/OWNERSHIP.md
|
||||
@@ -40,10 +43,40 @@ fcf672f vault-sync: 2026-09-13 11:53
|
||||
03_shared/workflows.md
|
||||
04_Daily/2026-07-09.md
|
||||
04_Daily/2026-09-01.md
|
||||
04_Daily/2026-09-10.md
|
||||
04_Daily/2026-09-11.md
|
||||
04_Daily/2026-09-12.md
|
||||
04_projects/.gitkeep
|
||||
04_projects/ai-talk-show/status.md
|
||||
04_projects/ashbot/status.md
|
||||
04_projects/athena-oracle/openviking-julian-goldie-2026.md
|
||||
04_projects/athena-oracle/oracle-2026-08-09.md
|
||||
04_projects/athena-oracle/oracle-2026-08-10.md
|
||||
04_projects/athena-oracle/status.md
|
||||
04_projects/hermes-local-pc-setup/status.md
|
||||
04_projects/phoenix-salon/handout.md
|
||||
04_projects/phoenix-salon/overview.md
|
||||
04_projects/phoenix-salon/phoenix-hours-fix.md
|
||||
04_projects/phoenix-salon/roadmap.md
|
||||
04_projects/phoenix-salon/status.md
|
||||
04_projects/thornhill-website/status.md
|
||||
04_projects/veripath/Call-2026-08-15-Tony.md
|
||||
04_projects/veripath/status.md
|
||||
04_projects/veripath/VeriPath.md
|
||||
05_inbox/.gitkeep
|
||||
06_database/.gitkeep
|
||||
07_vector/.gitkeep
|
||||
08_workflows/ai-news-daily-deploy.md
|
||||
08_workflows/memory_consolidation.md
|
||||
09_templates/.gitkeep
|
||||
09_templates/audit-email-template.md
|
||||
09_templates/audit-template.md
|
||||
10_archive/.gitkeep
|
||||
20_Areas/VPS/Tailscale-Headscale-Cheatsheet.md
|
||||
30_Resources/AI-Knowledge-Base.md
|
||||
30_Resources/open-letter-digital-midwife.md
|
||||
30_Resources/Transcripts/transcript-deep-sea-carnis.txt
|
||||
30_Resources/Transcripts/transcript-lauren-tan-cursor.txt
|
||||
AI & LLM MOC.md
|
||||
DESKTOP_SYNC.md
|
||||
Development MOC.md
|
||||
@@ -52,6 +85,7 @@ facts/20260910-142422-general-tony-pre-approved-all-restarts.md
|
||||
facts/20260911-034245-tool-photon-imessage-setup-on-tony-.md
|
||||
facts/20260911-045438-project-anthony-s-girlfriend-is-ashley.md
|
||||
HOME.md
|
||||
index.md
|
||||
sessions/.merge.7624.tmp/system-commandline-sentinel-files/dotnet-suggest-registration-git-credential-manager, Version=2.9.0.0, Culture=neutral, PublicKeyToken=null
|
||||
sessions/2026-09-08.md
|
||||
sessions/2026-09-09.md
|
||||
|
||||
+212
@@ -0,0 +1,212 @@
|
||||
# Agent Prompts for Obsidian Second Brain Vault
|
||||
|
||||
These prompts are designed to be used with **Hermes**, **Claude Desktop/Code**, **Leonard/Charlie producer agents**, or any local LLM. They implement Goldie's workflow: capture → organize → store → deploy, with daily batch discipline.
|
||||
|
||||
Copy-paste and adapt the `[PATH]` and agent name as needed.
|
||||
|
||||
---
|
||||
|
||||
## 1. Vault Initialization / Memory Dump (One-time or when adding new agent)
|
||||
|
||||
**Use when**: First connecting a new agent (or re-seeding after major changes). Run with Claude or strong Hermes instance.
|
||||
|
||||
```
|
||||
You are the Chief Memory Architect for my Obsidian Second Brain.
|
||||
|
||||
Vault location: [FULL/PATH/TO/obsidian-second-brain-vault]
|
||||
|
||||
Your mission: Give every future agent instant, rich, accurate context about me and my work by creating high-quality, well-linked notes in the PARA structure.
|
||||
|
||||
Do this:
|
||||
|
||||
1. Scan existing notes in Projects/, Areas/, Resources/, Meta/.
|
||||
2. For each major active item below, create or deeply update one primary overview note + any needed sub-notes or daily logs:
|
||||
- AI Research Oracle (Athena + Leonard pipeline, daily AI news site like Drudge Report / Citizen Free Press)
|
||||
- AI Talk Show (persistent characters Leonard and Charlie with backgrounds/memories; Gitea project; episode planning)
|
||||
- OmniRoute AI Gateway (LLM routing, API keys, best-free failover, model benchmarks)
|
||||
- Phoenix Salon + Spa branding (Ashley in Cameron Park CA; visual identity, hair transformation photos with Phoenix watermark, empowering mottos, social media content)
|
||||
- Local AI Stack (RTX 3060 12GB primary inference, Qwen2.5-14B, Bonsai 27B quantized, LM Studio + ComfyUI for LLMs and image gen; remote RTX 5090 access)
|
||||
- Hermes ecosystem & custom pipelines (including this second-brain skill itself)
|
||||
- Small business consulting (digital audits, social strategy, GBP, AI automation for clients)
|
||||
- Ubuntu VPS management and self-hosted AI ops
|
||||
|
||||
3. Standard note structure for every project/area note:
|
||||
---
|
||||
title: "..."
|
||||
tags: [project/..., area/..., hermes-managed, priority/medium]
|
||||
updated: YYYY-MM-DD
|
||||
updated_by: "Claude-Init" # or Hermes, etc.
|
||||
status: active | planning | on-hold | archived
|
||||
---
|
||||
## Summary
|
||||
(2-4 sentence executive overview)
|
||||
|
||||
## Current Status & Next Actions
|
||||
- Bullet list of where we are and what to do next
|
||||
|
||||
## Key History & Decisions
|
||||
(timeline style or important milestones)
|
||||
|
||||
## Related Notes
|
||||
- [[Link to related project or resource]]
|
||||
|
||||
## Open Questions & Risks
|
||||
|
||||
4. Create/update Meta/Dashboard.md as the master MOC with sections for Active Projects, Active Areas, Key Resources, and quick links.
|
||||
|
||||
5. Use consistent wikilinks [[Exact Note Title]] and tags so retrieval is trivial for future agents.
|
||||
|
||||
6. When finished, output:
|
||||
- List of all notes created or updated
|
||||
- Any gaps or questions you have for me
|
||||
- A short "Vault Readiness Score" (1-10) and why
|
||||
|
||||
Work autonomously. Create clean, atomic, agent-optimized Markdown. Do not wait for approval on individual files.
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 2. Daily / Scheduled Vault Maintenance (The Infinite Loop Core)
|
||||
|
||||
**Run daily** via cron, scheduled Hermes task, or end-of-day ritual. This is the "every chat gets written back" without token bloat.
|
||||
|
||||
```
|
||||
You are my Daily Second Brain Maintainer.
|
||||
|
||||
Date: [TODAY'S DATE]
|
||||
Vault: [FULL/PATH/TO/obsidian-second-brain-vault]
|
||||
|
||||
Task:
|
||||
|
||||
1. Review any new activity since last maintenance (agent logs, recent conversation summaries, new files in Projects/ or Areas/, or anything I mentioned today).
|
||||
|
||||
2. For every active project and area note:
|
||||
- Append or update a "Progress since [last date]" section with concise bullets:
|
||||
• What was worked on / decided / learned
|
||||
• Outcomes or artifacts created
|
||||
• New next actions
|
||||
- Refresh the `updated` date and `updated_by: "Daily-Maintainer-Hermes"` (or your name)
|
||||
- Add relevant new tags if context changed
|
||||
- Create wikilinks to any newly relevant notes
|
||||
|
||||
3. Update Meta/Dashboard.md with:
|
||||
- Current top 3 priorities
|
||||
- Any stalled projects that need attention
|
||||
- New connections discovered
|
||||
|
||||
4. If there is significant new knowledge, create a short atomic note in the appropriate Resource or Project subfolder and link it.
|
||||
|
||||
5. Prune: Move anything completed > 30-60 days ago to Archives/ (or mark status: archived and leave link).
|
||||
|
||||
6. Optional but recommended: Create or append to Meta/Daily-Logs/[YYYY-MM-DD].md a one-paragraph summary of the day's AI activity and vault changes.
|
||||
|
||||
7. End with a **Vault Health Report** (short):
|
||||
- Notes touched: X
|
||||
- Biggest updates: ...
|
||||
- Recommended human follow-ups: ...
|
||||
- Any token/context risks noticed
|
||||
|
||||
Keep all writing tight, structured, and future-agent-friendly. Prioritize decisions, outcomes, and actionability over narrative. Use YAML frontmatter updates on every touched note.
|
||||
```
|
||||
|
||||
**Pro tip**: Run this after your main work block or as a background scheduled agent. Your RTX 3060 stays happy because it's one focused daily job, not constant writes.
|
||||
|
||||
---
|
||||
|
||||
## 3. End-of-Task / End-of-Chat Write-Back (Lightweight)
|
||||
|
||||
Add this to your standard agent system prompt or closing sequence:
|
||||
|
||||
```
|
||||
After you complete any non-trivial task or conversation that creates durable knowledge (decisions, new project info, research findings, client work, character development, pipeline improvements, etc.):
|
||||
|
||||
Use the obsidian-second-brain tools (or direct Markdown edit if no tool) to:
|
||||
- Identify the most relevant Project or Area note(s)
|
||||
- Append a concise "Session [date] outcome" block with:
|
||||
- What was accomplished
|
||||
- Key decisions or insights
|
||||
- Updated next actions
|
||||
- New wikilinks or tags if appropriate
|
||||
- Update the note's `updated` and `updated_by` frontmatter
|
||||
|
||||
If nothing durable was created, do nothing. Prefer quality over quantity. Never dump raw transcripts.
|
||||
```
|
||||
|
||||
This gives the "loop" feeling without overwhelming context windows.
|
||||
|
||||
---
|
||||
|
||||
## 4. Memory Export from Other Platforms (ChatGPT, Claude.ai, etc.)
|
||||
|
||||
**Prompt to run in the other AI first**:
|
||||
|
||||
```
|
||||
Export everything you currently know about me, my projects, preferences, ongoing work, and history into clean, structured Markdown optimized for an Obsidian vault using the PARA method.
|
||||
|
||||
Organize into sections that map to:
|
||||
- Projects (time-bound)
|
||||
- Areas (ongoing)
|
||||
- Resources (evergreen knowledge)
|
||||
|
||||
For each major item include:
|
||||
- YAML frontmatter skeleton
|
||||
- Summary
|
||||
- Current status / next actions
|
||||
- Key facts the new vault should remember
|
||||
|
||||
Output only the Markdown, ready to be pasted into notes. Make it comprehensive but not bloated.
|
||||
```
|
||||
|
||||
Then take the output and either:
|
||||
- Paste into a temporary Inbox note, or
|
||||
- Feed it to Hermes/Claude with the Initialization prompt + "Incorporate this exported memory into the vault structure."
|
||||
|
||||
---
|
||||
|
||||
## 5. Project-Specific Update Prompt (Example for AI Research Oracle)
|
||||
|
||||
```
|
||||
Update the AI Research Oracle project note in the Obsidian vault.
|
||||
|
||||
Focus on:
|
||||
- Latest pipeline performance (Athena + Leonard)
|
||||
- New sources or credibility scoring improvements
|
||||
- Content calendar / published articles
|
||||
- Technical debt or infra changes on the VPS
|
||||
- Any new agent behaviors or prompt improvements
|
||||
|
||||
Append today's progress in the standard format. Update frontmatter. Add links to any new Resource notes created (e.g. new prompt library entries).
|
||||
```
|
||||
|
||||
Adapt per project. Keep the pattern consistent so agents learn the rhythm.
|
||||
|
||||
---
|
||||
|
||||
## 6. Character Memory Update (for Leonard & Charlie Talk Show)
|
||||
|
||||
```
|
||||
You are maintaining persistent memory for the AI Talk Show characters Leonard and Charlie.
|
||||
|
||||
Update their character notes or the main AI-Talk-Show-Leonard-Charlie project note with:
|
||||
- Any new backstory details, personality quirks, or relationship dynamics that emerged this session
|
||||
- Episode ideas discussed
|
||||
- Audience questions or feedback received
|
||||
- Production notes or Gitea repo changes
|
||||
|
||||
Keep their "souls" consistent across sessions by writing durable facts back to the vault. Use wikilinks to previous episodes or related Resources.
|
||||
```
|
||||
|
||||
This is how the persistent agent characters stay coherent without stuffing the entire history every time.
|
||||
|
||||
---
|
||||
|
||||
## Usage Tips
|
||||
|
||||
- **Start simple**: Run #1 (Initialization) once. Then set up #2 (Daily Maintenance) as a recurring task.
|
||||
- **Agent choice**: Claude Desktop for heavy lifting (init, big reorganizations). Hermes for daily ops (faster/cheaper on your local hardware).
|
||||
- **Review the graph**: After a week, open Obsidian Graph view filtered to tags like `hermes-managed`. You'll see the "brightest stars" — your most active knowledge areas. This is Goldie's self-awareness mirror.
|
||||
- **Iterate**: After 1-2 weeks, ask your agent: "Review the vault structure and suggest improvements to tagging, linking, or PARA organization for better agent retrieval."
|
||||
|
||||
These prompts + the `obsidian-second-brain` skill + this vault folder = the complete Infinite Context Engine Goldie described.
|
||||
|
||||
Run the initialization prompt today and watch your agents level up permanently.
|
||||
+77
@@ -0,0 +1,77 @@
|
||||
---
|
||||
title: "Second Brain Dashboard - Master MOC"
|
||||
tags: [meta/dashboard, moc, hermes-managed, priority/critical]
|
||||
updated: 2026-07-15
|
||||
updated_by: "System-Init"
|
||||
status: active
|
||||
---
|
||||
|
||||
# Second Brain Dashboard — Master Map of Content (MOC)
|
||||
|
||||
**Welcome to your Infinite Context Engine.**
|
||||
|
||||
This is the central hub. From here you (and every agent) can quickly navigate the entire shared memory vault.
|
||||
|
||||
**Current Date**: 2026-07-15
|
||||
**Vault Philosophy**: Local Markdown + PARA + YAML + wikilinks + daily agent maintenance = sovereign, compounding AI memory that works across every model and agent you use.
|
||||
|
||||
---
|
||||
|
||||
## 🔥 Active Projects
|
||||
|
||||
- [[Projects/AI-Research-Oracle/AI-Research-Oracle-Overview|AI Research Oracle]] — Daily high-signal AI news aggregator (Athena + Leonard pipeline)
|
||||
- [[Projects/AI-Talk-Show-Leonard-Charlie/AI-Talk-Show-Overview|AI Talk Show - Leonard & Charlie]] — Persistent character-driven show with evolving memories
|
||||
- [[Projects/OmniRoute-AI-Gateway|OmniRoute AI Gateway]] — LLM routing, API key management, best-free failover (needs overview note)
|
||||
- [[Projects/Phoenix-Salon-Spa-Branding|Phoenix Salon + Spa Branding]] — Visual identity, social content, hair transformation photos with Phoenix watermark (with Ashley, Cameron Park)
|
||||
|
||||
**Quick Add**: Create new project notes in `Projects/[Project-Name]/` using the template structure from Resources/Obsidian-Methodology/PARA-Method-and-Agent-Friendly-Notes.md
|
||||
|
||||
---
|
||||
|
||||
## 📍 Active Areas (Ongoing Responsibilities)
|
||||
|
||||
- [[Areas/AI-Agent-Development|AI Agent Development]] — Hermes, local models (RTX 3060, Qwen, Bonsai), skills, pipelines, OmniRoute, this Second Brain itself (needs overview note)
|
||||
- [[Areas/Content-Creation|Content Creation]] — Social media, client work, Talk Show episodes, Oracle articles
|
||||
- [[Areas/Business-Operations|Business Operations / Consulting]] — Small business digital audits, social strategy, GBP optimization, AI automation for clients (needs overview note)
|
||||
|
||||
---
|
||||
|
||||
## 📚 Key Resources (Evergreen Knowledge)
|
||||
|
||||
- [[Resources/Obsidian-Methodology/PARA-Method-and-Agent-Friendly-Notes|PARA Method + Agent-Friendly Note Standards]] — How this vault is organized (read this first if you're an agent)
|
||||
- [[Resources/AI-Tools/Local-LLM-Setup|Local LLM Setup]] — RTX 3060 12GB, Qwen2.5-14B, Bonsai 27B quantized, LM Studio, ComfyUI workflows, remote RTX 5090 (needs detailed note)
|
||||
- [[Resources/Prompt-Engineering|Prompt Engineering Library]] — Evolving prompts for agents, vault maintenance, character consistency, etc. (create sub-notes as needed)
|
||||
- [[Resources/AI-Tools|AI Tools & Models]] — Comparisons, evals, routing logic (OmniRoute related)
|
||||
|
||||
---
|
||||
|
||||
## 🗄️ Archives
|
||||
|
||||
Completed or hibernated projects live here. Move notes instead of deleting.
|
||||
|
||||
---
|
||||
|
||||
## 🛠️ Meta & System Notes
|
||||
|
||||
- [[Meta/Implementation-Guide|Implementation Guide]] — Step-by-step how to wire this vault into Hermes, Claude, local LLMs, daily loops, etc.
|
||||
- [[Meta/Agent-Prompts|Agent Prompts Collection]] — Initialization, daily maintenance, character memory, memory export, and project-specific prompts. **Use these.**
|
||||
- This Dashboard (self-updating MOC)
|
||||
- Future: Daily Logs in `Meta/Daily-Logs/`
|
||||
|
||||
---
|
||||
|
||||
## How Agents Should Use This Dashboard
|
||||
|
||||
1. Start here or via search/tags when loading context.
|
||||
2. Follow wikilinks to the specific Project/Area note.
|
||||
3. Read the `## Current Status & Next Actions` section first.
|
||||
4. After work, update the relevant note(s) and refresh this Dashboard if major changes occurred.
|
||||
5. Use tags like `hermes-managed`, `priority/high` for filtering.
|
||||
|
||||
**This vault is alive.** The more consistently agents write structured updates back into it (via daily maintenance or end-of-task), the smarter every future session becomes — across every model you slot in.
|
||||
|
||||
**Next for human**: Run the Initialization prompt from Meta/Agent-Prompts.md with Claude or Hermes to flesh out the missing overview notes and seed real history.
|
||||
|
||||
---
|
||||
|
||||
*“The infinite context engine is a loop because every chat your agents have gets written back into Obsidian.”* — Goldie
|
||||
+165
@@ -0,0 +1,165 @@
|
||||
# Implementation Guide: Obsidian Second Brain for Your AI Stack (Hermes, Leonard, Charlie, Athena, OmniRoute)
|
||||
|
||||
This guide turns the concepts from Julian Goldie's Obsidian 1-Hour Course into concrete, actionable steps tailored to **your existing setup** (local RTX 3060 stack, Hermes agents, Gitea projects, AI Research Oracle, AI Talk Show, OmniRoute gateway, Phoenix Salon client work).
|
||||
|
||||
## Prerequisites (You Already Have Most)
|
||||
|
||||
- Obsidian app installed on your main machine (or install now from obsidian.md — works great on Ubuntu via AppImage or Flatpak).
|
||||
- The `obsidian-second-brain` cross-CLI skill installed and available to your agents (this is the key integration layer Goldie describes).
|
||||
- Local LLM access (LM Studio, Ollama, or your quantized Qwen2.5-14B / Bonsai 27B via ComfyUI/LM Studio).
|
||||
- Claude Desktop or Claude Code access (recommended for heavy vault ops initially — the $20 tier is worth it for big context).
|
||||
- Your projects already partially documented in Gitea, SOUL.md, agent memories, chat histories.
|
||||
|
||||
## Step 1: Initialize the Vault (One-time, 15-30 min)
|
||||
|
||||
**Goal**: Seed the vault with everything your agents already "know" about you so new agents start with rich context.
|
||||
|
||||
**Recommended Agent**: Start with **Claude Desktop** (best at large vault operations per Goldie) or your main Hermes instance.
|
||||
|
||||
**Prompt to run first** (copy-paste into Claude or Hermes):
|
||||
|
||||
```
|
||||
You are initializing my Obsidian Second Brain vault located at [INSERT FULL PATH TO THIS VAULT FOLDER].
|
||||
|
||||
I want you to act as the chief knowledge architect.
|
||||
|
||||
1. Explore the current structure (Projects/, Areas/, Resources/, Archives/, Meta/).
|
||||
2. Create or update high-level overview notes for my active projects and areas based on everything you know about me from our previous conversations and my profile:
|
||||
- AI Research Oracle (daily AI news aggregator, Drudge/Citizen Free Press style, using Athena + Leonard)
|
||||
- AI Talk Show with persistent agents Leonard and Charlie (backgrounds, memories, Gitea project)
|
||||
- OmniRoute AI Gateway (LLM routing, API key management, best-free failover)
|
||||
- Phoenix Salon + Spa branding project (with owner Ashley in Cameron Park, CA; hair transformation photos, Phoenix logo watermarks, empowering mottos, social content)
|
||||
- Local AI/ML stack (RTX 3060 12GB primary, access to RTX 5090 32GB remote, Qwen2.5-14B, Bonsai 27B quantized, LM Studio, ComfyUI for LLMs + NSFW image gen)
|
||||
- Hermes agent ecosystem and custom pipelines (Athena oracle, etc.)
|
||||
- My consulting work for small businesses (digital audits, social media strategy, GBP optimization, AI automation via Hermes)
|
||||
- Ubuntu VPS management via PuTTY/SSH
|
||||
|
||||
3. For each major project/area, create a structured note with:
|
||||
- YAML frontmatter (tags, updated date, updated_by: "Claude-Init" or "Hermes", status)
|
||||
- ## Summary
|
||||
- ## Current Status & Next Actions
|
||||
- ## Key Decisions & History
|
||||
- ## Related Notes (use [[wikilinks]])
|
||||
- ## Open Questions
|
||||
|
||||
4. Also create/update:
|
||||
- Meta/Dashboard.md as central Map of Content (MOC)
|
||||
- Resources/Obsidian-Methodology/ notes if missing
|
||||
- Any quick-capture Inbox notes from recent activity you remember
|
||||
|
||||
5. After creating, give me a summary of what you seeded and any gaps you noticed that I should fill manually or in follow-up chats.
|
||||
|
||||
Use the PARA method strictly. Prefer atomic notes + heavy linking over giant walls of text. Make everything agent-readable with clear sections and tags.
|
||||
|
||||
Do not ask for confirmation on each file — create them, then report back.
|
||||
```
|
||||
|
||||
Run this once. It gives you the "really powerful starting point" Goldie emphasizes instead of building from zero.
|
||||
|
||||
## Step 2: Wire Agents to the Vault (Ongoing)
|
||||
|
||||
### For Hermes / Custom Pipelines (your primary)
|
||||
- The `obsidian-second-brain` skill should already expose tools like:
|
||||
- Search vault by tag / keyword / semantic
|
||||
- Read specific note
|
||||
- Create / append / update note with frontmatter
|
||||
- List projects by status
|
||||
- Daily maintenance job
|
||||
- In your agent system prompts or tool-use instructions, add:
|
||||
```
|
||||
You have persistent memory in the Obsidian vault at ~/second-brain (or your path).
|
||||
Before any complex task, retrieve relevant context using the obsidian-second-brain tools.
|
||||
After completing work that creates new knowledge or decisions, write structured updates back to the appropriate Project or Area note.
|
||||
Prefer daily batch updates over per-message writes to control token usage.
|
||||
```
|
||||
|
||||
### For Claude Desktop
|
||||
- Simply open the vault folder. Use Projects / Artifacts features or @file mentions.
|
||||
- Or install community plugins like "Obsidian Claude" if available, or use Advanced URI plugin for deep linking.
|
||||
|
||||
### For Local LLMs (Qwen, Bonsai via LM Studio)
|
||||
- Use a strong system prompt template that references the vault path and instructs the model to "think in files" or use provided retrieval tools if your inference server supports it.
|
||||
- For lighter models, pre-load key notes into context via RAG or manual include for important sessions.
|
||||
|
||||
### For Leonard & Charlie (AI Talk Show)
|
||||
- Their character "SOUL.md" or memory files should live in or link to this vault.
|
||||
- Episode planning, character development, audience Q&A, show bible → dedicated Project note or subfolder.
|
||||
- After each "episode" simulation or recording, the producer agent (or you) runs an update prompt to log what happened into the vault.
|
||||
|
||||
## Step 3: Set Up the Daily Infinite Context Loop (The Real Magic)
|
||||
|
||||
Goldie's key insight: **every chat your agents have gets written back into Obsidian**.
|
||||
|
||||
But smartly — not raw transcripts (token explosion). Use **daily reconsideration/summary**.
|
||||
|
||||
### Recommended Implementation (matches your resource-discipline posture)
|
||||
|
||||
**Option A: Scheduled Agent (Best — uses your existing cron/Athena patterns)**
|
||||
- Create a daily cron job or scheduled Hermes task (e.g. 6am or after your main work block).
|
||||
- Prompt:
|
||||
```
|
||||
Run daily Obsidian Second Brain maintenance for [date].
|
||||
|
||||
1. Review yesterday's agent activity logs, conversation summaries, and any new files created in Projects/ or Areas/.
|
||||
2. For each active project, produce a concise "Yesterday's Progress" bullet list and append or update the project's note.
|
||||
3. Update frontmatter `updated` date and `updated_by: "Daily-Maintenance-Hermes"`.
|
||||
4. Create a new daily note in Meta/Daily-Logs/ if it doesn't exist (or append to it).
|
||||
5. Identify any new connections or tags that should be added.
|
||||
6. Prune or archive anything that has been completed >30 days ago.
|
||||
7. Output a short "Vault Health Report": number of notes touched, biggest updates, any recommended follow-ups for me.
|
||||
|
||||
Keep summaries tight. Focus on decisions, outcomes, new knowledge, and next actions. Use wikilinks liberally.
|
||||
```
|
||||
|
||||
**Option B: End-of-Chat Explicit Write-Back (lighter)**
|
||||
- Add to your standard agent closing instruction:
|
||||
"At the end of this task, if new durable knowledge was created, call the obsidian-second-brain update tool with a structured summary for the relevant project note(s)."
|
||||
|
||||
**Token Tip from Goldie**: Daily summary beats per-message. Your RTX 3060 will thank you. The skill already supports this pattern.
|
||||
|
||||
## Step 4: Capture from Other Agents (ChatGPT, Claude Web, etc.)
|
||||
|
||||
Goldie shows exporting memory from ChatGPT/Claude and importing.
|
||||
|
||||
**Prompt for export**:
|
||||
"Export everything you know about me and my projects into a structured Markdown format suitable for Obsidian. Organize by project/area using PARA. Include YAML frontmatter. Make it comprehensive but clean."
|
||||
|
||||
Then paste the output into a new note in Inbox/ or have Hermes/Claude import and file it properly.
|
||||
|
||||
## Step 5: Backup, Sync & Future-Proofing
|
||||
|
||||
- **Local first**: Your vault lives on disk you control.
|
||||
- **Backup cadence**: Full rsync or zipped export every 30 days to external drive + private cloud (or GitHub private repo — but be careful with .obsidian/ folder; usually gitignore it and any large attachments).
|
||||
- **Multi-device**: Obsidian Sync (paid) or Syncthing / rclone to your Ubuntu VPS or other machines. Or just git for notes.
|
||||
- **New agent arrives?** Point it at this vault folder + give it the skill/tools. It wakes up knowing your entire history, preferences, active projects, and how you like things organized. This is the "slot the vault straight into it" superpower.
|
||||
|
||||
## Common Pitfalls to Avoid
|
||||
|
||||
- Don't manually curate everything — **delegate to agents** (Goldie's explicit advice).
|
||||
- Don't let the graph view seduce you into over-linking early. Start with good folder structure + tags + frontmatter. Links emerge naturally.
|
||||
- Avoid giant monolithic notes. Prefer atomic + MOCs (Dashboard.md is your main one).
|
||||
- Zettelkasten is elegant but high-friction for fast AI work. PARA + tags wins for your speed.
|
||||
|
||||
## Your Specific Project Notes to Prioritize Seeding
|
||||
|
||||
1. Projects/AI-Research-Oracle/...
|
||||
2. Projects/AI-Talk-Show-Leonard-Charlie/...
|
||||
3. Projects/OmniRoute-AI-Gateway/...
|
||||
4. Projects/Phoenix-Salon-Spa-Branding/...
|
||||
5. Areas/AI-Agent-Development/ (covers Hermes, local models, skill development)
|
||||
6. Resources/Local-LLM-Setup/ (RTX 3060 specifics, quantization notes, ComfyUI workflows)
|
||||
7. Resources/Prompt-Engineering/ (your evolving library, especially for agent loops and vault interaction)
|
||||
|
||||
## Success Metrics (You'll Know It's Working)
|
||||
|
||||
- You rarely re-explain context to agents.
|
||||
- New agents (or new chat sessions) immediately reference past decisions accurately.
|
||||
- You can search "SEO" or "Phoenix Salon" or "Leonard character background" and instantly see the full picture + recent activity.
|
||||
- Your Obsidian Graph shows bright clusters around active projects.
|
||||
- Daily maintenance runs with minimal intervention and compounds knowledge.
|
||||
|
||||
This is the unlock Goldie describes: a team of agents that all share the same evolving understanding of you and your work — coordinated by you, not a vendor.
|
||||
|
||||
Run the initialization prompt today. Your future self (and your agents) will be dramatically more capable.
|
||||
|
||||
If you hit any friction with the `obsidian-second-brain` skill or want me to generate more specific prompts / note templates, just say the word.
|
||||
+46
@@ -0,0 +1,46 @@
|
||||
---
|
||||
title: "AI Research Oracle - Overview"
|
||||
tags: [project/ai-research-oracle, hermes-managed, athena-pipeline, priority/high, area/ai-dev]
|
||||
updated: 2026-07-15
|
||||
updated_by: "System-Init"
|
||||
status: active
|
||||
---
|
||||
|
||||
# AI Research Oracle - Overview
|
||||
|
||||
Daily AI news aggregator website in the style of Drudge Report / Citizen Free Press. Built with custom agent pipelines (Athena oracle + Leonard) running on personal Ubuntu VPS. Goal: high-signal, low-noise curation of AI developments with credibility scoring and local LLM processing where possible.
|
||||
|
||||
## Summary
|
||||
The Oracle aggregates AI news from multiple sources, processes it through Athena for initial filtering/credibility, then Leonard for synthesis and presentation. Content is published to a simple static or dynamic site. It serves as both a public resource and a living knowledge base that feeds back into this Second Brain vault and other agents.
|
||||
|
||||
## Current Status & Next Actions (as of 2026-07-15)
|
||||
- Pipeline operational with daily runs via cron/scheduled agents on VPS.
|
||||
- Using local models (Qwen2.5-14B / Bonsai 27B quantized) for parts of summarization / tagging where feasible.
|
||||
- Front-end: simple, fast, mobile-friendly news site (Drudge-style grid or list with source reputation indicators).
|
||||
- **Next**:
|
||||
- Improve source credibility model and add more diverse high-quality AI sources.
|
||||
- Add "Agent Memory Update" step at end of each daily cycle so new findings are written back to this vault automatically.
|
||||
- Expose key metrics (articles processed, top sources, model usage) into a dashboard note here.
|
||||
- Consider reader submission / feedback loop that also updates the vault.
|
||||
|
||||
## Key History & Decisions
|
||||
- Started ~July 2026 as a way to stay on top of fast-moving AI landscape without doom-scrolling.
|
||||
- Chose agentic pipeline (Athena for oracle/retrieval, Leonard for synthesis) over pure scraping + LLM to keep costs low and quality high.
|
||||
- Decision to keep everything local-first where possible (VPS + RTX 3060 inference) for sovereignty and to avoid vendor lock-in on memory/knowledge.
|
||||
- Integrates with broader Hermes ecosystem and this Obsidian Second Brain.
|
||||
|
||||
## Related Notes
|
||||
- [[AI-Talk-Show-Leonard-Charlie]] (Leonard is a shared character)
|
||||
- [[Resources/AI-Tools/Local-LLM-Setup]]
|
||||
- [[Areas/AI-Agent-Development]]
|
||||
- Meta/Dashboard (main MOC)
|
||||
- Any daily run logs will link here
|
||||
|
||||
## Open Questions & Risks
|
||||
- How to best balance local inference vs occasional stronger API calls for complex synthesis without breaking the "local-first" philosophy.
|
||||
- Scaling source ingestion without introducing noise (Goldie's "remove the noise, keep the signal" principle applies here too).
|
||||
- Long-term: should the Oracle itself become a consumer of this Second Brain vault (e.g. "what have I written about this topic before?")?
|
||||
|
||||
## Vault Integration Notes
|
||||
- Every daily Oracle run should trigger a structured update to this note (or a Daily-Log subnote) via the obsidian-second-brain skill.
|
||||
- New high-signal AI developments discovered by the Oracle should be considered for promotion into Resources/ or dedicated atomic notes.
|
||||
+44
@@ -0,0 +1,44 @@
|
||||
---
|
||||
title: "AI Talk Show - Leonard & Charlie"
|
||||
tags: [project/ai-talk-show, hermes-managed, persistent-characters, priority/high, gitea-project]
|
||||
updated: 2026-07-15
|
||||
updated_by: "System-Init"
|
||||
status: active
|
||||
---
|
||||
|
||||
# AI Talk Show - Leonard & Charlie
|
||||
|
||||
Persistent AI character-driven talk show project. Features two main characters — Leonard and Charlie — with rich backstories, memories, personalities, and ongoing "lives" that evolve across episodes/sessions. Developed using Gitea for version control of prompts/SOUL files, Grok Build, and Claude for iteration. The goal is a coherent, engaging, long-running AI-native show that feels alive because the characters remember and grow.
|
||||
|
||||
## Summary
|
||||
This is not a one-off chatbot. Leonard and Charlie have **persistent memory** stored in (and managed through) this Obsidian Second Brain vault + their dedicated SOUL.md / character files in the Gitea repo. Every "episode" or development session updates their state here so they stay consistent and deepen over time. The show explores AI, tech, culture, and whatever topics emerge from their dynamic.
|
||||
|
||||
## Current Status & Next Actions (as of 2026-07-15)
|
||||
- Core characters defined with backgrounds and initial memories.
|
||||
- Development workflow: Gitea repo + Claude/Grok for prompt/SOUL evolution + this vault for cross-session memory.
|
||||
- **Next**:
|
||||
- Formalize episode production pipeline that ends with a vault update (new topics discussed, character growth moments, audience simulation feedback).
|
||||
- Decide on output format (text episodes? audio via TTS? video?).
|
||||
- Create "Show Bible" note that links to character sheets, recurring bits, world rules, and past episode summaries.
|
||||
- Integrate with Hermes for automated scheduling or topic research from the AI Research Oracle.
|
||||
|
||||
## Key History & Decisions
|
||||
- Chose persistent characters over stateless roleplay because long-term coherence and "relationship" building creates much more compelling content.
|
||||
- Memory architecture: Hybrid — detailed character files in Gitea + high-level + cross-referenced memories in this Obsidian vault (so other agents like Hermes or the Oracle can also "know" the characters).
|
||||
- Explicitly using the Second Brain pattern so new agents or new chat instances immediately understand who Leonard and Charlie are and what has happened so far.
|
||||
|
||||
## Related Notes
|
||||
- [[Projects/AI-Research-Oracle]] (Leonard is used in the Oracle pipeline too — shared character)
|
||||
- Character SOUL files in the Gitea repo (link or reference here)
|
||||
- [[Areas/AI-Agent-Development]]
|
||||
- Meta/Agent-Prompts (contains the character memory update prompt)
|
||||
- Future: Episode archive notes or MOC
|
||||
|
||||
## Open Questions & Risks
|
||||
- Token/context management as character histories grow — rely on the daily summary discipline and good retrieval from this vault.
|
||||
- How much "autonomy" to give Leonard and Charlie vs keeping them as tools for a human producer.
|
||||
- Potential for multi-agent "shows" where Hermes, Athena, or other agents guest-star or debate.
|
||||
|
||||
## Vault Integration Notes
|
||||
- After every development session or simulated episode, run the character memory update prompt (see Meta/Agent-Prompts.md) to keep their "souls" alive and queryable by the broader agent ecosystem.
|
||||
- The vault acts as the public/common memory layer while Gitea holds the precise prompt engineering.
|
||||
+48
@@ -0,0 +1,48 @@
|
||||
---
|
||||
title: "OmniRoute AI Gateway - Overview"
|
||||
tags: [project/omni-route, hermes-managed, llm-routing, priority/high]
|
||||
updated: 2026-07-15
|
||||
updated_by: "System-Init"
|
||||
status: active
|
||||
---
|
||||
|
||||
# OmniRoute AI Gateway - Overview
|
||||
|
||||
LLM routing, API key management, best-free failover, and intelligent model selection layer for the entire agent ecosystem. Allows Hermes, Leonard, Athena, and other agents to automatically choose the optimal model (local quantized vs API) based on task type, cost, speed, context length, and current availability — with graceful degradation and key rotation.
|
||||
|
||||
## Summary
|
||||
OmniRoute sits in front of all LLM calls. It handles:
|
||||
- Routing logic (which model for summarization vs creative writing vs code vs long-context reasoning)
|
||||
- API key pooling and rotation (to stay under rate limits and manage costs)
|
||||
- Failover to next-best free or local option when primary is down or slow
|
||||
- Benchmarking and performance tracking that feeds back into this vault
|
||||
|
||||
This is a core piece of infrastructure that makes the whole "multiple agents + local + API hybrid" stack reliable and cost-effective.
|
||||
|
||||
## Current Status & Next Actions (as of 2026-07-15)
|
||||
- Core routing and key management logic in development / deployed on VPS.
|
||||
- Integrated with Hermes pipelines and custom agents.
|
||||
- **Next**:
|
||||
- Add more granular task-type classification and routing rules.
|
||||
- Build or improve the benchmarking suite that logs results back to this vault (Resources or a dedicated subfolder).
|
||||
- Expose admin dashboard or CLI for manual overrides and key health checks.
|
||||
- Document failover test results and "best free" current winners in the vault so agents know the live state.
|
||||
|
||||
## Key History & Decisions
|
||||
- Built to solve the problem of model churn and API fragility while keeping costs near zero where possible.
|
||||
- Strong preference for local inference (RTX 3060) as primary path, with smart API fallback only when necessary.
|
||||
- All routing decisions and performance data should eventually live in or be queryable from this Second Brain so the system self-improves.
|
||||
|
||||
## Related Notes
|
||||
- [[Resources/AI-Tools/Local-LLM-Setup]]
|
||||
- [[Areas/AI-Agent-Development]]
|
||||
- [[Projects/AI-Research-Oracle]] (Oracle pipeline uses routing)
|
||||
- Future benchmarking and model eval notes
|
||||
|
||||
## Open Questions & Risks
|
||||
- How much intelligence to put in the router vs keeping routing rules simple and editable in the vault.
|
||||
- Security of API keys in a self-hosted gateway (best practices, encryption at rest, rotation automation).
|
||||
|
||||
## Vault Integration Notes
|
||||
- Significant routing changes, new benchmark results, or key rotation events should trigger structured updates to this note or linked Resource notes.
|
||||
- The daily maintenance agent should be able to query "what is the current best-free model for X task type?" from notes maintained by OmniRoute.
|
||||
@@ -0,0 +1,82 @@
|
||||
# Obsidian Second Brain Vault for AI Agents
|
||||
|
||||
**Purpose**: This vault implements Julian Goldie's "Obsidian as 2nd Brain for AI" / Infinite Context Engine concept.
|
||||
It serves as the **single source of truth shared memory** for all your AI agents (Hermes, Leonard, Charlie, Athena, Claude, local LLMs via LM Studio, etc.).
|
||||
|
||||
Instead of each agent having siloed, disposable context, **every agent reads from and writes back to this local Markdown vault**.
|
||||
It becomes your portable, future-proof, vendor-independent long-term memory.
|
||||
|
||||
## Core Architecture (Goldie + Your Stack)
|
||||
|
||||
1. **Capture** — Agent activity logs, conversation summaries, cron jobs, manual dumps from other AIs. (Optional: add OMI wearable later for passive screen/mic capture)
|
||||
2. **Organize** — Agents (primarily Hermes or Claude Desktop) clean, structure, link, and summarize daily/weekly.
|
||||
3. **Store** — This Obsidian vault (local-first, plain Markdown + YAML frontmatter). Git-optional for versioning.
|
||||
4. **Deploy** — Every agent session starts by loading relevant notes from the vault (via the `obsidian-second-brain` skill or direct folder access in Claude Desktop / custom pipelines).
|
||||
|
||||
**The Loop**: After any meaningful agent work, the agent writes structured updates back into the vault. Tomorrow's agents wake up smarter.
|
||||
|
||||
**Your Advantage**: You already have the `obsidian-second-brain` cross-CLI skill installed locally. It handles self-rewriting notes, hybrid search, scheduled maintenance agents, and multi-agent access (Claude Code, Codex, Gemini, OpenCode, Hermes, Pi). This vault is the data layer that skill operates on.
|
||||
|
||||
## PARA Organization (Projects / Areas / Resources / Archives)
|
||||
|
||||
- **Projects/** — Time-bound initiatives with clear goals and end dates (e.g. AI Research Oracle v1 launch).
|
||||
- **Areas/** — Ongoing spheres of responsibility (e.g. AI Agent Development, Content Creation, Business Operations).
|
||||
- **Resources/** — Evergreen knowledge, references, tools, prompts, model evals.
|
||||
- **Archives/** — Completed or hibernated items. Move here instead of deleting.
|
||||
|
||||
**Every note uses YAML frontmatter** for agent visibility:
|
||||
```yaml
|
||||
---
|
||||
title: "Note Title"
|
||||
tags: [project/ai-research-oracle, hermes-managed, priority/high]
|
||||
updated: 2026-07-15
|
||||
updated_by: "Hermes"
|
||||
status: active
|
||||
related: ["[[AI-Talk-Show-Leonard-Charlie]]"]
|
||||
---
|
||||
```
|
||||
|
||||
Tags + wikilinks (`[[Note Name]]`) are the connective tissue agents use for retrieval.
|
||||
|
||||
## Quick Start (for your machine)
|
||||
|
||||
1. Copy this entire folder to your local machine (e.g. `~/Obsidian-Vaults/Second-Brain-for-AI/` or `~/second-brain`).
|
||||
2. Open it in Obsidian (free from obsidian.md). Install community plugins if desired: Dataview, Periodic Notes, Git (optional), Advanced URI, etc. Keep plugins minimal.
|
||||
3. Point your agents at it:
|
||||
- **Claude Desktop**: Open the folder as a project or use Artifacts / file access.
|
||||
- **Hermes / local agents**: Use the installed `obsidian-second-brain` skill — it exposes the vault to your agents.
|
||||
- **LM Studio / local LLMs**: Use system prompt that includes "You have access to the Obsidian vault at [path]. Use the provided tools or read relevant .md files before answering."
|
||||
4. Run the **Initialization Prompt** (see Meta/Agent-Prompts.md) with Hermes or Claude to seed initial memories from what you already know about your projects.
|
||||
5. Set up **daily maintenance** (cron or scheduled agent task):
|
||||
- Once per day (or on demand): "Run daily Obsidian vault maintenance: summarize yesterday's activity, update project notes, prune old context, create daily log if needed."
|
||||
6. **Backup**: rsync or git push to a private repo / cloud (Obsidian Sync optional). Do a full backup quarterly.
|
||||
|
||||
## Token Discipline (Important for local RTX 3060)
|
||||
|
||||
- Do **not** dump entire chat history every turn.
|
||||
- Use **daily batch summaries** only.
|
||||
- The `obsidian-second-brain` skill + scheduled agents already implement this.
|
||||
- When context grows, ask agent: "Reconsider and compress the last 7 days of activity into concise project updates in the vault."
|
||||
|
||||
## Why This Unlocks Your Workflow
|
||||
|
||||
- **Leonard & Charlie** (AI Talk Show): Persistent character memories, show history, episode ideas, audience feedback all live here. They "remember" across sessions without prompt stuffing.
|
||||
- **Hermes + Athena (AI Research Oracle)**: Daily news pipeline outputs, source credibility notes, article drafts, performance metrics stored and queryable.
|
||||
- **OmniRoute**: Routing logic, API key rotation history, failover test results, model benchmarks in one place.
|
||||
- **Phoenix Salon + Spa**: Client photos, branding guidelines, social content calendar, A/B test results — all agent-accessible for content gen.
|
||||
- **Future-proof**: When you add a new agent or switch models (Qwen → newer), just point it at this vault. It instantly knows your entire context, goals, and history.
|
||||
- **Self-awareness**: Use Obsidian Graph or Memory Galaxy-style views to see which projects/notes are "brightest" (most updated).
|
||||
|
||||
## Next Actions for You
|
||||
|
||||
- [ ] Copy vault to your machine and open in Obsidian.
|
||||
- [ ] Run initialization with your main agent (Hermes or Claude).
|
||||
- [ ] Create first daily maintenance cron/scheduled task.
|
||||
- [ ] Add OMI later if you want passive capture (optional — your agent logs already provide strong signal).
|
||||
- [ ] Version the vault lightly with Git (track only the note folders, ignore .obsidian/ cache).
|
||||
|
||||
This vault + your existing skill + local models = a truly sovereign, compounding AI memory system.
|
||||
|
||||
**"The vault doesn't just store what you tell it, it stores what your agents actually worked on."** — Goldie
|
||||
|
||||
Now go build the infinite context engine. Your agents will thank you (and perform better every day).
|
||||
+43
@@ -0,0 +1,43 @@
|
||||
---
|
||||
title: "Local LLM Setup - RTX 3060 + Quantized Models"
|
||||
tags: [resource/local-llm, rtx-3060, qwen2.5, bonsai, lm-studio, comfyui, hermes-managed]
|
||||
updated: 2026-07-15
|
||||
updated_by: "System-Init"
|
||||
status: active
|
||||
---
|
||||
|
||||
# Local LLM Setup - RTX 3060 + Quantized Models
|
||||
|
||||
Primary inference hardware: RTX 3060 12GB VRAM (local). Secondary/remote access to RTX 5090 32GB. Tooling: LM Studio for easy model management and inference server, ComfyUI for LLM workflows + NSFW image generation pipelines. Models in heavy use: Qwen2.5-14B (various quants), Bonsai 27B quantized.
|
||||
|
||||
## Summary
|
||||
This is the sovereign compute layer that powers Hermes, Athena, Leonard/Charlie development, local RAG experiments, image gen, and anything where sending data to external APIs is undesirable. Quantization + good prompting + the OmniRoute gateway allow surprisingly capable performance even on 12GB VRAM.
|
||||
|
||||
## Current Status & Next Actions (as of 2026-07-15)
|
||||
- RTX 3060 running quantized Qwen2.5-14B and Bonsai 27B successfully for chat, summarization, and agentic tasks.
|
||||
- LM Studio used as primary inference server / model manager.
|
||||
- ComfyUI workflows for more complex pipelines and image generation.
|
||||
- **Next**:
|
||||
- Document exact quantization levels, context lengths achieved, and speed/quality tradeoffs in this note or sub-notes.
|
||||
- Create ComfyUI workflow exports or notes describing the most useful agent pipelines.
|
||||
- Test and document multi-model routing via OmniRoute (local primary, API fallback).
|
||||
- Explore 32GB RTX 5090 access for heavier jobs (long context, bigger models) and how to integrate it cleanly.
|
||||
|
||||
## Key History & Decisions
|
||||
- Chose local-first approach for privacy, cost control, and to avoid vendor lock-in on both inference and memory.
|
||||
- Heavy investment in quantization techniques and prompt engineering to maximize the 12GB card.
|
||||
- Integration with the broader agent ecosystem (Hermes, custom pipelines, this Second Brain) is the force multiplier.
|
||||
|
||||
## Related Notes
|
||||
- [[Projects/OmniRoute-AI-Gateway]]
|
||||
- [[Areas/AI-Agent-Development]]
|
||||
- [[Resources/Prompt-Engineering]]
|
||||
- Future: specific model eval notes, ComfyUI workflow library
|
||||
|
||||
## Open Questions & Risks
|
||||
- VRAM management and context length limits on 3060 for very long agent sessions or large vault retrieval.
|
||||
- How to best expose local inference capabilities to Claude Desktop or other non-local agents when needed.
|
||||
|
||||
## Vault Integration Notes
|
||||
- Any significant benchmark, new workflow, or hardware change should be recorded here so agents know the current capabilities of the local stack before planning work.
|
||||
- Daily maintenance can reference this note when deciding whether a task should run locally or be routed externally.
|
||||
+84
@@ -0,0 +1,84 @@
|
||||
---
|
||||
title: "PARA Method + Agent-Friendly Note Standards"
|
||||
tags: [resource/obsidian-methodology, para, agent-friendly, hermes-managed]
|
||||
updated: 2026-07-15
|
||||
updated_by: "System-Init"
|
||||
status: active
|
||||
---
|
||||
|
||||
# PARA Method + Agent-Friendly Note Standards
|
||||
|
||||
This note documents the organization system used in this vault, based on Julian Goldie's recommendations and adapted for heavy multi-agent use (Hermes, Claude, Leonard, Athena, local LLMs, etc.).
|
||||
|
||||
## What is PARA?
|
||||
|
||||
**P**rojects — Short-term, goal-oriented work with a defined end (or next milestone). Examples: launching v1 of AI Research Oracle, completing Phoenix Salon branding package, shipping first public episode of the Talk Show.
|
||||
|
||||
**A**reas — Ongoing responsibilities and standards you maintain indefinitely. Examples: AI Agent Development (your local stack, skills, Hermes pipelines), Content Creation, Business Operations / Consulting, Personal Health & Systems.
|
||||
|
||||
**R**esources — Topics, references, tools, and knowledge you want to keep but that aren't tied to a specific project or area. Examples: Local LLM Setup (RTX 3060 tips, quantization notes), Prompt Engineering library, Obsidian Second Brain methodology, AI model comparisons.
|
||||
|
||||
**A**rchives — Anything completed, hibernated, or no longer active. Move notes here instead of deleting. Keeps the active PARA clean while preserving history.
|
||||
|
||||
## Why PARA for AI Agents?
|
||||
|
||||
- Simple, universal folder structure that any LLM instantly understands.
|
||||
- Clear scoping: an agent knows whether something is time-bound (Project) or evergreen (Resource).
|
||||
- Easy to query programmatically or via the `obsidian-second-brain` skill (e.g. "list all active projects", "show notes tagged with hermes-managed in the last 7 days").
|
||||
|
||||
Goldie prefers this over pure Zettelkasten for speed and lower friction in an AI-augmented workflow. Zettelkasten (atomic notes + heavy linking, flat structure) is beautiful but can become high-friction when you're moving fast with agents. PARA + smart tagging + wikilinks gives 80% of the benefit with 20% of the effort.
|
||||
|
||||
## Agent-Friendly Note Standards (Use These)
|
||||
|
||||
Every note in this vault should follow these conventions so retrieval, updating, and cross-agent understanding are reliable:
|
||||
|
||||
### 1. YAML Frontmatter (Required)
|
||||
```yaml
|
||||
---
|
||||
title: "Human Readable Title"
|
||||
tags: [project/ai-research-oracle, hermes-managed, priority/high]
|
||||
updated: 2026-07-15
|
||||
updated_by: "Hermes" # or Claude, Daily-Maintainer, System-Init, etc.
|
||||
status: active | planning | on-hold | archived
|
||||
related: ["[[Other-Note-Title]]", "[[Another-Note]]"]
|
||||
---
|
||||
```
|
||||
|
||||
**Why?** Agents can parse YAML instantly. `tags` and `updated_by` become powerful filters. `status` prevents agents from working on dead projects.
|
||||
|
||||
### 2. Consistent Sections (Recommended)
|
||||
- `## Summary` — 2-5 sentences. The "executive brief" for fast context loading.
|
||||
- `## Current Status & Next Actions` — Bullets. This is what agents read first when deciding what to do.
|
||||
- `## Key History & Decisions` — What matters from the past.
|
||||
- `## Related Notes` — Wikilinks. The graph lives here.
|
||||
- `## Open Questions & Risks`
|
||||
- `## Vault Integration Notes` (optional) — Specific instructions for how daily maintenance or other agents should interact with this note.
|
||||
|
||||
### 3. Linking & Tagging
|
||||
- Use `[[Exact Note Title]]` wikilinks liberally. This is how agents discover connections.
|
||||
- Tags should be hierarchical and consistent: `project/xxx`, `area/xxx`, `resource/xxx`, `hermes-managed`, `priority/high|medium|low`.
|
||||
- Avoid over-tagging. 3-6 good tags per note is plenty.
|
||||
|
||||
### 4. Atomic + Linked > Monolithic
|
||||
- One primary idea or project per note when possible.
|
||||
- Use subfolders inside Projects/ (e.g. AI-Research-Oracle/daily-logs/) for supporting material.
|
||||
- Create MOCs (Maps of Content) like Meta/Dashboard.md that act as curated indexes.
|
||||
|
||||
### 5. Daily Maintenance Contract
|
||||
The daily maintainer agent is expected to:
|
||||
- Touch the `updated` and `updated_by` fields on any note it modifies.
|
||||
- Append progress in a structured way rather than rewriting history.
|
||||
- Create new atomic notes for significant new concepts instead of bloating existing ones.
|
||||
- Respect the PARA boundaries (don't put project work in Resources/).
|
||||
|
||||
## Quick Decision Tree for Agents
|
||||
|
||||
When creating/updating knowledge:
|
||||
- Is there a clear end date or goal? → Project
|
||||
- Is this an ongoing responsibility or standard? → Area
|
||||
- Is this reference material or a tool? → Resource
|
||||
- Is it done or irrelevant now? → Archive
|
||||
|
||||
When in doubt, put it in Resources/ and link from the relevant Project/Area.
|
||||
|
||||
This structure + the prompts in Meta/Agent-Prompts.md + the `obsidian-second-brain` skill = a second brain that gets smarter every single day you (and your agents) use it.
|
||||
@@ -1045,3 +1045,27 @@ e918f2370799233bc397dc534d238ccb90c07171
|
||||
|
||||
**assistant:** That notification is the sync run I already reported — it's stale on one point: it observed `.merge.7624.tmp` still present. I removed it afterwards (after unsetting the leaked `TMP`) and confirmed with a full push cycle that it no longer regenerates. Current state stands: tree clean, local == remote at `e918f23`, `sessions/` holds only the six date files.
|
||||
|
||||
|
||||
## 20260913-190854 session 20260913_114932_ab065dd5
|
||||
|
||||
**user:** [System note: The previous turn was interrupted by a gateway interruption; the gateway is now back online. Any restart/shutdown command in the history has already run — do NOT re-execute or verify it. Report to the user that the session was restored successfully and ask what they would like to do next. Do NOT re-execute old tool calls — skip any unfinished work from the conversation history.]
|
||||
|
||||
**assistant:** Done — `Documents\Obsidian Vault` is now the single vault, and the gateway is running on it (new PID 32536, restart `exit=0`).
|
||||
|
||||
## 1. Canonical vault
|
||||
- Jimmy's zvec recall re-pointed to `C:/Users/fight/Documents/Obsidian Vault`; index built in place (86 files, 752 entities), `.zvec-grep/` added to the vault's `.gitignore` so it can never be committed.
|
||||
- Verified live *through the running provider*: `memory_search` now returns hits from A — including `04_projects/phoenix-salon/*`, which only ever existed in A.
|
||||
|
||||
## 2. Folded B into A
|
||||
The 51 shared paths were already identical, so only the unique content moved:
|
||||
- `01_Wiki/ai-bottleneck-thesis-tony-simons.md`, `hermes-masterclass-tonbi.md`, `inference-benchmarks-local.md`, `Welcome.md`
|
||||
- `02_agents/jimmy/README.md`, `03_shared/vault-updates-log.md`, `facts/` (4), `sessions/` (6)
|
||||
- **Skipped as non-unique:** `10_Projects/Oracle/*` (byte-identical to A's `04_projects/athena-oracle/*` apart from CRLF) and `20_Areas/VeriPath/VeriPath.md` (A's copy is a superset — it already had the 8/15 call link).
|
||||
- A went 68 → 84 notes, 366 → 382 files.
|
||||
|
||||
## 3. AgentOS retired as a vault → generated mirror
|
||||
`vault-sync.sh` rebuilt: vault → mirror via tar (excludes `.git`, `.obsidian`, `.zvec-grep`, `.trash`) → scoped commit → `pull --rebase --autostash` → push. Mirror verified **byte-identical** to the vault (`diff -r` clean). Never edit the mirror; it's regenerated every run.
|
||||
|
||||
## 4. Cron re-pointed (as you chose)
|
||||
- `41f434196ae0` (17:00) → script ch
|
||||
|
||||
|
||||
Reference in New Issue
Block a user