Cursor Agent 10d6dfe701 fix(vertex_ai/agent_engine): concat multi-text parts; surface hard-stop finish reasons ... Address Greptile review on PR #27139: - Concat all text parts in a chunk instead of dropping all but the first. - Surface hard-stop finish reasons (SAFETY, MAX_TOKENS, RECITATION, BLOCKLIST, PROHIBITED_CONTENT, SPII) even when the chunk has no user-facing content, so safety blocks don't look like benign intermediate chunks. Plain STOP on content-empty chunks is still dropped (the #19121 fix). Co-authored-by: Mateo Wang <mateo-berri@users.noreply.github.com>		2026-05-12 16:04:07 +00:00
..
a2a_protocol/providers/pydantic_ai_agents	style: run black formatter on files from main merge	2026-04-17 13:02:59 -07:00
integrations/helicone	style: run black formatter on files from main merge	2026-04-17 13:02:59 -07:00
litellm_core_utils	test: create fresh ModelResponse per test to avoid shared mutable state	2026-03-03 14:54:11 -03:00
llms	fix(vertex_ai/agent_engine): concat multi-text parts; surface hard-stop finish reasons	2026-05-12 16:04:07 +00:00
proxy	harden cloud file compatibility path	2026-05-01 18:04:06 -07:00
test_batch_completion_models_all_responses.py	style: run black formatter on files from main merge	2026-04-17 13:02:59 -07:00
test_bedrock_extended_beta_models.py	feat(bedrock): add support for 4 new beta models (#20974)	2026-02-13 18:32:12 +05:30
test_bedrock_nemotron_super.py	Litellm ishaan april2 (#25113)	2026-04-04 12:31:49 -07:00
test_no_hardcoded_secrets.py	style: run black formatter on files from main merge	2026-04-17 13:02:59 -07:00
test_proxy_auth.py	feat(sdk): add proxy_auth for auto OAuth2/JWT token management (#20238)	2026-02-02 22:04:08 -08:00
test_router_retry_backoff_headers.py	fix(router): use current retry exception for retry backoff (#20725)	2026-02-10 22:44:25 -08:00
test_stream_chunk_builder_images.py	style: run black formatter on files from main merge	2026-04-17 13:02:59 -07:00