Dylan 063083af15 [prompts] Better user_instructions handling (#1836) ...

## Summary
Our recent change in #1737 can sometimes lead to the model confusing
AGENTS.md context as part of the message. But a little prompting and
formatting can help fix this!

## Testing
- Ran locally with a few different prompts to verify the model
behaves well.
- Updated unit tests

2025-08-04 18:55:57 -07:00

..

common

fix: create separate test_support crates to eliminate #[allow(dead_code)] (#1667)

2025-07-24 12:19:46 -07:00

fixtures

Improve SSE tests (#1546)

2025-07-12 16:53:55 -07:00

cli_responses_fixture.sse

Add CLI streaming integration tests (#1542)

2025-07-12 18:05:58 -07:00

cli_stream.rs

chore: refactor exec.rs: create separate seatbelt.rs and spawn.rs files (#1762)

2025-07-31 13:11:47 -07:00

client.rs

[prompts] Better user_instructions handling (#1836)

2025-08-04 18:55:57 -07:00

compact.rs

ci fix (#1782)

2025-08-01 09:17:13 -07:00

exec_stream_events.rs

collabse stdout and stderr delta events into one (#1787)

2025-08-01 14:00:19 -07:00

exec.rs

[sandbox] Filter out certain non-sandbox errors (#1804)

2025-08-03 13:05:48 -07:00

live_agent.rs

fix command duration display (#1806)

2025-08-03 11:33:44 -07:00

live_cli.rs

Disallow expect via lints (#865)

2025-05-12 08:45:46 -07:00

sandbox.rs

feat: make .git read-only within a writable root when using Seatbelt (#1765)

2025-08-01 16:11:24 -07:00

stream_no_completed.rs

chore: refactor exec.rs: create separate seatbelt.rs and spawn.rs files (#1762)

2025-07-31 13:11:47 -07:00