feat(agent): forward model settings and recover from thinking stalls [PC-4672] - #1017

Open

tudormatei1 wants to merge 8 commits into

mainfrom

feat/model-specific-settings

Commits on Aug 18, 2026

feat(agent): forward model settings and skip forced tool_choice under Claude thinking
tudormatei1
committed
fc47974
handle anthropic thinking edge case
tudormatei1
committed
81dedc5
router and llm node now support one llm call stall, meaning thinking block with response in text
tudormatei1
committed
6a0ab10
refactor
tudormatei1
committed
ae06dad
refactor
tudormatei1
committed
33b7c59
ensured PII doesnt get logged in telemetry
tudormatei1
committed
191ee03
addressed copilot comments
tudormatei1
committed
a0f190a
fix import after licensing merged into llm module
tudormatei1
committed
9b88d18