Commit Graph

57 Commits

Author SHA1 Message Date
Farukest
c33f8d381b fix: correct off-by-one in retry exhaustion checks
The retry exhaustion checks used > instead of >= to compare
retry_count against max_retries. Since the while loop condition is
retry_count < max_retries, the check retry_count > max_retries can
never be true inside the loop. When retries are exhausted, the loop
exits and falls through to response.choices[0] on an invalid response,
crashing with IndexError instead of returning a proper error.
2026-03-01 02:27:26 +03:00
0xbyt4
dfd50ceccd fix: preserve Gemini thought_signature in tool call messages
Gemini 3 thinking models attach extra_content with thought_signature
to function call responses. This must be echoed back on subsequent
API calls or the server rejects with a 400 error. The assistant
message builder was dropping this field, causing all Gemini 3 Flash/Pro
tool-calling flows to fail after the first function call.
2026-02-28 18:10:05 +03:00
teknium1
50cb4d5fc7 fix(agent): update error message for unsupported Anthropic API endpoints to clarify usage of OpenRouter 2026-02-27 23:23:31 -08:00
Teknium
2bc9508b7c Merge pull request #173 from adavyas/fix/anthropic-base-url-guard
fix(agent): fail fast on Anthropic native base URLs
2026-02-27 23:22:01 -08:00
teknium1
19f28a633a fix(agent): enhance 413 error handling and improve conversation history management in tests 2026-02-27 23:04:32 -08:00
adavyas
0c0a2eb0a2 fix(agent): fail fast on Anthropic native base URLs 2026-02-27 21:19:29 -08:00
0xbyt4
90ca2ae16b test: add unit tests for run_agent.py (AIAgent)
71 tests covering pure functions, state/structure methods, and
conversation loop pieces. OpenAI client and tool loading are mocked.
2026-02-26 16:15:04 +03:00