get_loose_ends: clarify that an in-flight todo (e.g. a running bash task) doesn't need an early ack
This commit is contained in:
parent
8c51e37804
commit
6f208c7301
1 changed files with 5 additions and 0 deletions
|
|
@ -350,6 +350,11 @@ impl AgentServer {
|
|||
passing its id + kind to `cancel_loose_end`. Empty result is reported clearly. Todos \
|
||||
cap at 40 rows; a trailer line says how many more are pending — clear the shown ones \
|
||||
with `mark_todos_done`, then call again for the rest.\n\
|
||||
A row that is still genuinely in flight (a running bash task, say) doesn't need an \
|
||||
early ack — only mark done the ones you've actually resolved. A real content change \
|
||||
(like a task finishing) un-acks and resurfaces the same row on its own, so acking it \
|
||||
while it's still running buys nothing and just costs you a second, redundant call \
|
||||
later.\n\
|
||||
Pass `agent: \"<name>\"` to inspect a specific peer agent's threads. Direct \
|
||||
child agents are always accessible. For non-children, the `query_agent_state` \
|
||||
capability is required — without it the request is rejected with an error."
|
||||
|
|
|
|||
Loading…
Reference in a new issue