v2.0.144
·
12 commits
to master
since this release
v2.0.144 - WebFetch completed document handling
This release fixes the lab-gated native WebFetch path when the language server
has already executed read_url_content and returned a real web_document.
- Completed
read_url_content.web_documentsteps are no longer surfaced to
OpenAI clients as dead tool-call proposals. The proxy now preserves Cascade's
final assistant text withfinish_reason="stop". - If Cascade returns a completed WebFetch document but no final text, the proxy
falls back to the fetched document body as assistant content. - Pending WebFetch permission steps without a document keep the existing
proposal/wait behavior, and Bash/Read/Grep native proposal semantics are
unchanged. - Proto trace classification now checks
completed_web_documentbefore
pending_permission, so steps that contain bothweb_documentand a
requested-interaction echo are classified as completed.