Send parallel_tool_calls true on Codex Responses requests - #969
Merged
Merged
Conversation
A live probe against the ChatGPT Codex backend accepts parallel_tool_calls:true (HTTP 200 on POST /codex/responses, echoed true in response.created), so the host rewrites the field the packaged adapter still pins to false. Related to CL-7420.
Collaborator
Author
|
Audit: Approve. The parallel_tool_calls:true rewrite is scoped to the Codex Responses wire body with live-probe evidence cited (HTTP 200, echoed true, CL-7420) and the packaged-adapter false pin called out inline, so the next adapter bump has a clear conflict marker. Test asserts the on-the-wire value and IMPLEMENTATION.md is updated in the same change. Note: if the vendor adapter ever flips its pin to true, withParallelToolCalls becomes a pass-through and should be removed then. |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Probe + verdict for CL-7420: does the ChatGPT Codex backend accept parallel_tool_calls true?
Live probe (single trivial turn, Reply with exactly: ok, via codex/abk-labs, model gpt-5.5, POST https://chatgpt.com/backend-api/codex/responses with parallel_tool_calls flipped to true): HTTP 200 accepted. The backend echoed parallel_tool_calls:true in the response.created SSE event.
Verdict: accepted, so this flips the wire value to true. Omission (the OpenAI/Grok shape) was not probed — per the one-probe budget, true is the conservative choice because it is the only alternative with live evidence. The packaged @corbits/codex-provider adapter still pins false, so the host rewrites the field on the wire in src/provider/codex-responses.ts (request-body only; no store/continuity/reasoning, runner, gate, or TUI changes). Unit test updated to pin true; docs/IMPLEMENTATION.md updated in the same change.
No merges requested (review only).
Related to CL-7420.