feat(skippy): serve non-chat GGUF workloads via OpenAI API - #1833

Open

IvGolovach wants to merge 16 commits into

Mesh-LLM:mainfrom

IvGolovach:codex/non-chat-model-workloads

Commits on Sep 12, 2026

feat(skippy): serve and certify non-chat GGUF workloads

IvGolovach
authored and
codex
committed
c66892e
test(plugin): publish discard event after its reason
codex
committed
19f139d
fix(openai): gate workloads on each serving target
codex
committed
25ff652
fix(openai): preserve workload routing and multipart contracts
IvGolovach
committed
0bbc85b
ci(skippy): provision source-bound CPU workload canaries
IvGolovach
committed
d97715e
fix(skippy): allow safe serving listener restarts on Unix
IvGolovach
committed
b28ac30
test(skippy): align multimodal oracle execution contracts
IvGolovach
committed
95f301b

Commits on Sep 13, 2026

chore(ci): refresh serving console-print source location
IvGolovach
committed
b8584ca
fix(skippy): satisfy strict PCM conversion lint

scama
authored and
IvGolovach
committed
5934708
fix(skippy): close non-chat certification gaps

scama
authored and
IvGolovach
committed
e84efa7
fix(skippy): harden workload admission and certification
IvGolovach
committed
6681ca9
fix: reconcile workload serving with runtime event updates
IvGolovach
committed
821762c
test: require integer indexes in embedding certification
IvGolovach
committed
551c8d9
fix: reject invalid workload routes and certification evidence
IvGolovach
committed
ca0ef0f
fix(ci): reconcile workload review with current main
IvGolovach
committed
4cda960

Commits on Sep 14, 2026

fix(skippy): reconcile workload certification with current main
IvGolovach
committed
cadf698