What & why
The DTE series 5/5 feature (orchestrator new_task thinking_effort pass-through, #1332) needs end-to-end coverage at the real VS Code extension-host boundary: the wire-level proof that a child task started via new_task carries the requested effort in its first provider request, and that unsupported selections are rejected at the wire boundary without blocking delegation.
Unit tests cannot represent this failure mode (they do not exercise the extension host, the wire envelope, or the aimock replay/record round-trip), so the coverage lives in a standalone e2e addendum PR stacked on the feature head (kept off the feature PR to protect the 1000-line standalone diff cap).
Scope
apps/vscode-e2e suite for orchestrator new_task thinking_effort: effort pass-through, inheritance of the parent's effective effort, and negative rejection at the wire boundary.- Ephemeral effort proxy + aimock fixtures; deterministic key-free replay/record.
Feature issue: #1332 (DTE series 5/5). Related: #1328, #1329, #1330, #1331.
What & why
The DTE series 5/5 feature (orchestrator
new_taskthinking_effortpass-through, #1332) needs end-to-end coverage at the real VS Code extension-host boundary: the wire-level proof that a child task started vianew_taskcarries the requested effort in its first provider request, and that unsupported selections are rejected at the wire boundary without blocking delegation.Unit tests cannot represent this failure mode (they do not exercise the extension host, the wire envelope, or the aimock replay/record round-trip), so the coverage lives in a standalone e2e addendum PR stacked on the feature head (kept off the feature PR to protect the 1000-line standalone diff cap).
Scope
apps/vscode-e2esuite for orchestratornew_taskthinking_effort: effort pass-through, inheritance of the parent's effective effort, and negative rejection at the wire boundary.Feature issue: #1332 (DTE series 5/5). Related: #1328, #1329, #1330, #1331.