prompt() computed currentImages but never passed them to _queueSteer()
or _queueFollowUp() in the streaming branch. Both methods only accepted
text and built content as [{ type: 'text', text }], dropping images.
- _queueSteer/_queueFollowUp now accept optional ImageContent[]
- streaming branch in prompt() passes currentImages through
- public steer()/followUp() accept and forward optional images
- RPC types, handler, and client updated for steer/follow_up images
- rpc.md: document images on steer/follow_up, fix ImageContent examples
HTML div elements collapse whitespace by default, stripping leading
spaces from ANSI-rendered tool output (e.g. JSON code blocks).
Added white-space: pre-wrap to .ansi-line class.
update(source) unconditionally called updateSourceForScope for both
user and project scopes, creating a local install even when the source
was only registered globally. Unify into a single code path that
iterates settings per scope, filtering by identity when a source is
provided.
Tests used setExtensionPaths but update() iterates packages, not
extensions. All 5 non-pinned tests were silently passing as no-ops.
Also add regression test for scope-aware update behavior.
When creating a new session, initial model and thinking level were set
on the agent but never saved to session file. This caused --resume to
default thinking level to 'off'.
fixes#342
Aliou Diallo