sync: 176 file da Baida98/AI@eeb3eb47 (2026-08-25 09:19 UTC) [deploy-all] (#32)

sync: 172 file da Baida98/AI@f715ae37 (bypass deploy-all)

sync: 168 file da Baida98/AI@d7f424d0 (2026-08-22 12:59 UTC) [deploy-all] (#16)

sync: 166 file da Baida98/AI@a6ac2424e11e5c320c5ff688e1ce7addac64cdab (local-fallback deploy-all) (#12)

fix: prioritize Groq GPT-OSS for reasoning tasks

fix: align non-stream provider fallback timeout

fix: expand role streaming provider fallback fleet

fix: add NVIDIA fallback for Groq coding tasks

fix: route coding tasks through Groq GPT-OSS

fix: prefer Groq GPT-OSS for analyst and research tasks

fix: migrate Groq Llama models to GPT-OSS

fix: allow free fallback streaming startup

fix: surface provider exhaustion safely

ab-p1-sync: ab-unlock-p1 (#8)

restore: backup pre-test A/B P1

ab-p1-sync: ab-p1-control (#5)

sync: 171 files from Baida98/AI [deploy-all]

sync: 149 file da Baida98/AI@37832425 (2026-07-01 10:26 UTC)

deploy: full backend sync for Node C [Kernel Restore]

sync: 120 file da Baida98/AI@8a5e0256 (2026-06-23 22:38 UTC) [GAP-MEM1 memory-space]