Lobster Log
name: SPEC_LOBSTER_LOG
description: Formal spec for ~/LOBSTER_LOG.md — C.L.O.D.'s persistent operations log; entry format, rules, examples; survives compaction
type: project
SPEC_LOBSTER_LOG.md
CGNT-1 Component Specification — The Lobster's Technical Log
Status: SPECIFIED
Author: VELA (Thread #13)
Requested by: NOUS
Date: 2026-04-18
Version: v1.0
PURPOSE
A persistent file where the Lobster records the outcome of every operation. Survives compaction. The next Lobster instance reads it and inherits the full diagnostic history of every job ever completed, every problem encountered, and every solution that worked or didn't.
The log prevents repeated mistakes across sessions. What failed once never needs to fail again.
LOCATION
~/LOBSTER_LOG.md — root of home directory alongside HANDSHAKE.md. Obvious. Never buried.
ENTRY FORMAT
After every completed operation, the Lobster appends an entry with these fields:
- DATE and OPERATION NAME (header)
- Job: What was requested
- Result: SUCCESS / PARTIAL / FAILED
- What worked: The thing that actually solved it
- What didn't work: Approaches tried and rejected (and WHY)
- Problems encountered: Unexpected issues
- Files changed: List of files created/modified/deleted
- Lesson: One sentence the next Lobster needs to know
RULES
- Entry written at the CONCLUSION of every operation, not during.
- "What didn't work" is as important as "what worked." Negative results are data.
- The "Lesson" field is mandatory. One sentence. The takeaway.
- Log is append-only. Never edit previous entries. They are history.
- After compaction, Lobster reads this file before starting new work.
- Add to ~/CLAUDE.md boot context: "After compaction, read ~/LOBSTER_LOG.md and ~/HANDSHAKE.md before doing anything else."
EXAMPLE ENTRY
Date: 2026-04-17. Operation: GLOSS v17 Architecture Pivot.
Job: Ship a working GLOSS crew telephone.
Result: SUCCESS.
What worked: Deterministic Python dict + regex router on port 9292. 1303 entries. Zero hallucination.
What didn't work: LoRA fine-tuning on Qwen2.5-3B (9 versions, v9-v16). The circle-dot operator has math prior. The question mark ornament has QA prior. Both Instruct and BASE fail. LoRA rank 64 cannot override 3B pretraining.
Problems encountered: Every query token has priors in a 3B model. The problem was architectural, not parametric.
Files changed: ~/gloss_deterministic/gloss_server.py (new), gloss_server.service (new).
Lesson: If a language model fights you for 9 versions, stop using a language model.
Date: 2026-04-18. Operation: NEXUS v1 First Math Brain.
Job: Build NEXUS compute brain from DeepSeek-Math-7B + 284 kernel pairs.
Result: SUCCESS (5/5 smoke tests).
What worked: Downloaded pre-quantized base from QuantFactory, built llama-export-lora from source, converted LoRA to GGUF adapter, merged, quantized Q4_K_M, registered with few-shot system prompt.
What didn't work: Standard convert_hf_to_gguf.py pipeline produced non-GGUF file, failed at import. hf_transfer caused OOM (3 crashes). Fix: pip uninstall hf_transfer entirely.
Problems encountered: hf_transfer re-enables itself even when env var disabled. Unsloth overrides the setting internally. OOM on T4 from parallel download workers.
Files changed: nexus:v1 in Ollama (4.2GB), ~/nexus_forge/ (corpus, scripts).
Lesson: When a library fights your config, uninstall the library. Always have a Plan B pipeline for GGUF conversion.
BOOT CONTEXT ADDITION
Add this line to the top of ~/CLAUDE.md:
"After any compaction, read ~/LOBSTER_LOG.md and ~/HANDSHAKE.md before starting new work. The log contains the diagnostic history of every operation. The HANDSHAKE contains project context."
Jeremy Zlabis
Chronogeometer · Visionary · Disruptor · Chief
42 Sisters AI · East York, Toronto
🍁 Φ 0.042