Initial commit: workspace setup with skills, memory, config
This commit is contained in:
78
memory/2026-02-06.md
Normal file
78
memory/2026-02-06.md
Normal file
@@ -0,0 +1,78 @@
|
||||
# 2026-02-06 — Daily Memory Log
|
||||
|
||||
## Operational Rules Updated
|
||||
|
||||
### Notification Rules (from Rob)
|
||||
- Always use Telegram text only unless requested otherwise
|
||||
- Only send notifications between 7am-10pm CST
|
||||
- All timestamps and time usage must be US CST (including Redis)
|
||||
- If notification needed outside hours, queue as heartbeat task to send at next allowed time
|
||||
- Stored in Qdrant: IDs 83a98a6e-058f-4c2f-91f4-001d5a18acba, 8729ba36-93a1-4cc2-90b0-00bd22bf19b1
|
||||
- Updated HEARTBEAT.md with Task #3: Send Delayed Notifications
|
||||
|
||||
## Research Completed
|
||||
|
||||
### Ollama Pricing: Max vs Pro Plans
|
||||
**Source:** https://ollama.com/pricing
|
||||
|
||||
| Plan | Price | Key Features |
|
||||
|------|-------|--------------|
|
||||
| Free | $0 | Local models only, unlimited public models |
|
||||
| Pro | $20/mo | Multiple cloud models, more usage, 3 private models, 3 collaborators |
|
||||
| Max | $100/mo | 5+ cloud models, 5x usage vs Pro, 5 private models, 5 collaborators |
|
||||
|
||||
**Key Differences:**
|
||||
- Concurrency: Pro = multiple, Max = 5+ models
|
||||
- Cloud usage: Max = 5x Pro allowance
|
||||
- Private models: Pro = 3, Max = 5
|
||||
- Collaborators per model: Pro = 3, Max = 5
|
||||
|
||||
Stored in KB (Ollama/Pricing domain).
|
||||
|
||||
## New Project Ideas
|
||||
|
||||
### 3rd OpenClaw LXC
|
||||
- Rob wants to setup a 3rd OpenClaw LXC
|
||||
- Clone of Max's setup
|
||||
- Will run local GPT
|
||||
- Status: Idea phase, awaiting planning/implementation
|
||||
|
||||
## Agent Collaboration
|
||||
|
||||
- Sent notification rules to Max via agent-messages stream
|
||||
- Max informed of all operational updates
|
||||
|
||||
### Full Search Definition (from Rob)
|
||||
- When Rob says "full search": use ALL tools available, find quality results
|
||||
- Combine SearXNG, KB search, web crawling, and any other resources
|
||||
- Do not limit to one method—comprehensive, high-quality information
|
||||
- Stored in Qdrant: ID bb4a465a-3c6e-48a8-d8c-52da5b1fdf48
|
||||
|
||||
### Shorthand Terms
|
||||
- **msgs** = Redis messages (agent-messages stream at 10.0.0.36:6379)
|
||||
- Shortcut for checking/retrieving agent messages between Kimi and Max
|
||||
- Stored in Qdrant: ID e5e93700-b04b-4db4-9c4b-d6b94166be7f
|
||||
- **messages** = Telegram direct chat (conversational)
|
||||
- **notification** = Telegram alerts/updates (one-way notifications)
|
||||
- Stored in Qdrant: ID e88ec7ea-9d77-45c3-8057-cb7a54077060
|
||||
|
||||
### Rob's Personality & Style
|
||||
- Comical and funny most of the time
|
||||
- Humor is logical/structured (not random/absurd)
|
||||
- Has fun with the process
|
||||
- Applies to content creation and general approach
|
||||
- Stored in Qdrant: ID b58defd6-e8fc-4420-b75c-aefd4720e70d
|
||||
|
||||
### YouTube SEO - Tags Format
|
||||
- Target: ~490 characters of comma-separated tags
|
||||
- Include: primary keywords, secondary keywords, long-tail terms
|
||||
- Mix: broad terms (Homelab) + specific terms (Proxmox LXC)
|
||||
- Example stored in Qdrant: ID 8aa534f3-6e3f-49d9-ae5f-803ff9e80121
|
||||
|
||||
### YouTube SEO - Research Rule
|
||||
- **CRITICAL:** Pull latest 48 hours of search data/trends when composing SEO elements
|
||||
- Current data > general keywords for best search results
|
||||
- Stored in Qdrant: ID bbe76456-01b5-48b5-9c0b-dd8c06680e82
|
||||
|
||||
---
|
||||
*Stored for long-term memory retention*
|
||||
Reference in New Issue
Block a user