Cross-AI · Research Desk
Research Desk
Sequential multi-runtime research package. Date 2026-09-10 JST. RESULT PASS. No Library Verified claim.
ChatGPT was NOT used in the verified run.
No ChatGPT session, seat, or synthesis label. The final brief is labeled
executor-synthesis — an honest runtime name, not ChatGPT.
Where it sits
Cross-AI orchestration package with public proof. Uses a Library Ollama recipe (botshelf-deep-research-3b) as one seat. It is not a Marketplace listing and not a Verified Library badge.
Runtimes that actually ran (sequential, not simultaneous)
| # | Runtime | What happened |
|---|---|---|
| 1 | web-fetch-curl | 8 public Library URLs · HTTP 200 |
| 2 | ollama:botshelf-deep-research-3b | ~13.8s · llama3.2 3.2B |
| 3 | local-file-audit | Proof Engine local files audit |
| 4 | executor-synthesis | Final brief · NOT ChatGPT |
Skipped / not claimed: ChatGPT (unavailable) · Claude as invent seat · X discovery (UNAVAILABLE).
Job + outcome
Library Verified vs Untested policy + Proof Engine monitoring vs data-analysis distribution.
RESULT PASS — 2026-09-10 JST. Conflicts recorded; sequential envelopes retained in internal QC (redacted public proof below).
Limitations
- Ollama soft invent risk — low-confidence open questions on Verified criteria are acceptable; do not treat model speculation as policy.
- Synthesis prefers HTML + PE — public Library HTML and Proof Engine files outrank Ollama guesses when they conflict.
- No Verified claim — this run does not flip any Library badge.