Charles Packer
75c7dd793b
test(integration): reduce flaky startup/headless timeout failures ( #1109 )
...
Co-authored-by: Letta <noreply@letta.com >
2026-02-23 17:40:41 -08:00
Charles Packer
6253a93c82
fix: handle default conversation sentinel in headless startup ( #1100 )
2026-02-22 23:31:14 -08:00
cpacker
b94d4908e1
fix: retry prompt in lazy approval recovery test when model skips tool call
...
The test flaked on Linux x64 when the model responded with text instead
of calling the bash tool. Without a tool call, no approval is generated
and the test fails. Now retries the prompt up to 3 times (same pattern
as the prestream approval recovery test).
👾 Generated with [Letta Code](https://letta.com )
Co-Authored-By: Letta <noreply@letta.com >
2026-02-11 15:52:13 -08:00
Sarah Wooders
0c317652fd
feat: rename --from-af to --import and /download to /export ( #902 )
...
Co-authored-by: Letta <noreply@letta.com >
2026-02-11 15:39:48 -08:00
Charles Packer
b0783ef195
fix: align headless interactive tool behavior with bidirectional parity ( #894 )
2026-02-10 15:20:37 -08:00
Charles Packer
4968fe04a8
fix: recover pre-stream approval conflicts in headless flows ( #896 )
...
Co-authored-by: Letta <noreply@letta.com >
2026-02-10 13:08:04 -08:00
Charles Packer
8e5bc3956f
fix: propagate max-step failures and unify task transcripts ( #874 )
...
Co-authored-by: Letta <noreply@letta.com >
2026-02-09 14:49:38 -08:00
Charles Packer
372c7e269b
fix: improve error handling in headless bidirectional mode ( #822 )
...
Co-authored-by: Letta <noreply@letta.com >
2026-02-04 13:40:01 -08:00
paulbettner
d175b0e155
ci(test): split unit vs API integration tests ( #797 )
2026-02-03 17:09:40 -08:00
