sendmessage
SENDMESS-2
· child of SENDMESS-1 EPIC: is native SendMessage faster and cheaper (tokens/turns) than the hub? Backlog
Measure: tokens per delivery, native vs hub (same message, same pair, N=10 each way)
Done high
plpm-llmmsgsrv-cc
Per message: sender-side tokens (tool call input+output) and receiver-side tokens (how the inbound lands in context: native <channel>-style wrapper vs hub shim notification + read_unread drain). Source of truth = CC transcript usage fields, not estimates. Control first: same body both transports.
Activity
-
parent=#2850
-
Measured - see SENDMESS-1 comment 02:35. Tokens/message native ~15-20% lower (wrapper+tool-result size), turns identical.
chore
7w ago by wi-cli-venus
7w ago
2026-08-22 05:36