Repository navigation
Conversation
…#183, #184) #182 Skill tool crashed with "replace() argument 2 must be str, not dict" when a model passed args as an object. Coerce the unambiguous shapes (null, single-string-value object, list of strings, number) and return an error naming the expected shape for anything else. A non-string name gets the same treatment. #183 _fetch_custom_model_limit never hit its cache for "custom/<model>" and never remembered a miss, so /v1/models was fetched before every model and tool call. Look the model up as given, then with a leading "custom/" removed (not bare_model(), which would also strip the vendor segment of an already-bare "Qwen/Qwen3-Coder"). Remember "endpoint doesn't describe this model" for the process and a failed request for 30s. #184 Skill returned the sub-run's whole transcript, including reasoning chunks and agent status lines. Return only the final assistant turn, with streamed text as the fallback when the run died before completing a turn. The tool output cap now receives config, so a user-set context_window and the custom endpoint's per-model limit apply.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
…, #183, #184)
#182 Skill tool crashed with "replace() argument 2 must be str, not dict" when a model passed args as an object. Coerce the unambiguous shapes (null, single-string-value object, list of strings, number) and return an error naming the expected shape for anything else. A non-string name gets the same treatment.
#183 _fetch_custom_model_limit never hit its cache for "custom/" and never remembered a miss, so /v1/models was fetched before every model and tool call. Look the model up as given, then with a leading "custom/" removed (not bare_model(), which would also strip the vendor segment of an already-bare "Qwen/Qwen3-Coder"). Remember "endpoint doesn't describe this model" for the process and a failed request for 30s.
#184 Skill returned the sub-run's whole transcript, including reasoning chunks and agent status lines. Return only the final assistant turn, with streamed text as the fallback when the run died before completing a turn. The tool output cap now receives config, so a user-set context_window and the custom endpoint's per-model limit apply.