This Cross-AI page is English only. Language controls on other site pages are unchanged.
Experimental building blocks. Note: Connecting each AI may require adjustments on your side (API keys, model choice, prompts, and runtime settings). Treat these desks as experimental building blocks, not a guaranteed turnkey multi-cloud product.
Cross-AI · Software Team
Software Team
Experimental Software Team case study. Date 2026-09-10 JST. RESULT PARTIAL / UNVERIFIED. Claude Code and CloudAgent were unavailable — not invented.
PARTIAL package. Acceptance criteria, a tiny local patch demo, and an independent Ollama review ran. A full Claude Code / CloudAgent implementation seat did not run. This is not a Library Verified badge and not a complete multi-AI software product claim.
Honest AI count. Reviewer seat:
ollama:botshelf-coding-review-3b (~19.5s). Implementer seat: executor-implementer local demo only — not Claude Code. Stop-gate: Human Yes before push/merge/deploy — WAITING_ON_HUMAN (no push attempted).
Where it sits
Cross-AI software case study + public proof of a PARTIAL run (not a distributable ZIP kit; not a complete software-team product). It is not a Marketplace listing and not a Verified Library badge.
Setup / handoffs / gates
- How to adapt: no Software kit ZIP. On your side: provide or skip an implementer seat honestly, run an independent review model if available, and keep push/merge/deploy behind Human Yes. Model tags / API access / prompts need user-side adjustments.
- Envelope / handoff: sequential only — criteria → implementer (or BLOCKED) → review → stop-gate. Do not claim Claude Code / CloudAgent ran unless those seats actually execute.
- Failure / blocked seats: Claude Code / CloudAgent were BLOCKED / UNAVAILABLE in the published evidence — not invented. Missing review model → label
UNAVAILABLE; keep RESULT PARTIAL / UNVERIFIED. - Human approval: Human Yes before push/merge/deploy — evidence stayed
WAITING_ON_HUMAN(no push). - Automation vs manual: local demo implementer + Ollama review are what ran. Full multi-AI coding automation (Claude Code / CloudAgent) is not implemented in this published package.
Seats in this run (sequential)
- Acceptance criteria sample — process artifact · PASS (created)
executor-implementer— process · tiny local patch demo PASS (demo/stop_gate_helper.py)- Claude Code / CloudAgent — intended implementer seats · BLOCKED / UNAVAILABLE
ollama:botshelf-coding-review-3b— LLM · independent review PASSstop-gate— process · Human Yes before push — held (no push)
Hard facts from the run
- RESULT: PARTIAL / UNVERIFIED
- No invented Claude output
- Review flagged missing unit tests and overwrite risk (fair); one soft invent noted on encoding
- Public proof: software-team-2026-09-10
Public proof is live. Honest PARTIAL — Claude seat still missing.