feat(models): codex second opinions default to gpt-6-astra - #24
Merged
Merged
Conversation
…gpt-6-astra
Add an optional default_model to the Advisor spec and a per-advisor model
alias map. The codex advisor now defaults its second-opinion model to
gpt-6-astra (GPT-6 Astra), with gpt6/astra as case-insensitive aliases
scoped to codex so a name never leaks across CLIs.
- effective_model() resolves the model once and is used by build_command,
the session registry record, and command.json, so the resolved id (not a
raw alias or empty string) is what gets persisted.
- default_model is never injected on resume, avoiding a mid-thread model
switch on stored codex threads.
- Escape hatches: advisors.json ("codex": {"default_model": null}) or
--model default to suppress the flag entirely.
Docs and CHANGELOG updated; no version bump.
# Conflicts: # tests/test_cli.py
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Adds GPT-6 Astra (
gpt-6-astra, released 2026-09-03) as the default second-opinion model for the codex advisor, with a clean opt-out. Codex-native model, so it's scoped to codex only.What changed
Advisor.default_model— optional per-advisor default model. Onlycodexsets it (gpt-6-astra); every other advisor is unchanged.MODEL_ALIASES,resolve_model) —gpt6/astra→gpt-6-astra, case-insensitive, scoped to codex so a shorthand never leaks to another CLI. Deliberately nofablealias (the Claude CLI resolves that itself).effective_model()— single source of truth. The resolved id (not a raw alias or empty string) is what gets sent on argv and persisted to the session registry andcommand.json.default_modelis never injected on a resume; explicit--modelstill applies exactly as before.advisors.json"codex": {"default_model": null}, or--model defaultto suppress the flag entirely.Behaviour change
Codex second opinions with no
--modelnow send--model gpt-6-astra. All other advisors are byte-for-byte unchanged. Documented above and in the CHANGELOG.Testing
python -m pytest -q→ 335 passed (was 298; +37 incl. 20 new tests).main).--model defaultsentinel, resume skips the default, resolved id persisted to registry/command.json,"default_model": nulloverride.Plan (with the pre-implementation review notes addressed) is in
docs/plans/gpt6-astra-default-model.md.