The tasks build the inference path: emsha-backed SHA-256, inferproxy,
config, a hand-written HTTP and SSE client, request building, delta
assembly, the chat state machine, the thinking cap, the slot gate with
retry, the startup self-test and on-device verification.
Everything the tasks copy in was checked against a private reference
implementation: the gate passes after each task in order, the timing
tests pass repeatedly under CPU load, and the reference passes the
self-test and all four device checks on straylight. Expected results
for the recorded streams were derived by a separate script.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A request pinned to a busy slot receives no bytes until the slot frees.
reasoning_control ends a thinking block on demand; the capped turn is
re-read once on the next request.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
spike/m0.py is throwaway. Findings are in docs/inference-contract.md:
cache reuse and tool parsing pass through chat-completions, the tools
array must stay fixed per epoch, and the shared router cannot meet the
slot contract. Clean throughput and slot pinning are still open because
another session was using the GPU.
docs/decisions.md lists the brief changes this implies as proposals.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
kyleandClaude Fable 5.1