PR #23175: fix(cli): drop ChatGPT-account-rejected models from openai-codex picker

• Repository: NousResearch/hermes-agent
• Author: konsisumer
• State: closed | merged: False
• Link: https://github.com/NousResearch/hermes-agent/pull/23175

Description (problem / solution / changelog)

Codex /model picker no longer offers gpt-5.2-codex, gpt-5.1-codex-max, or gpt-5.1-codex-mini in its offline fallback — the ChatGPT-account Codex backend rejects them with HTTP 400.

What changed and why

• hermes_cli/codex_models.py: removed gpt-5.2-codex, gpt-5.1-codex-max, and gpt-5.1-codex-mini from DEFAULT_CODEX_MODELS. The openai-codex provider always authenticates through the ChatGPT-account OAuth backend (resolve_codex_runtime_credentials hard-codes auth_mode="chatgpt" and points at chatgpt.com/backend-api/codex), which rejects those slugs with HTTP 400: "The '<model>' model is not supported when using Codex with a ChatGPT account." Whenever live model discovery isn't available (offline first run, transient API failure), the picker fell back to this list and let the user pick a model the backend would reject on the very next request.
• Comment block updated to record why the three slugs are intentionally absent and to point at the issue. They remain in _PROVIDER_MODELS["openai"] (api.openai.com) and the opencode-zen catalog because those backends still accept them.
• tests/hermes_cli/test_codex_models.py: added test_default_codex_models_excludes_chatgpt_account_unsupported asserting the three rejected slugs stay out of DEFAULT_CODEX_MODELS.

Forward-compat synthesis (_FORWARD_COMPAT_TEMPLATE_MODELS) is unchanged: gpt-5.4, gpt-5.4-mini, and gpt-5.3-codex-spark are still synthesized from gpt-5.3-codex (which stays in the fallback), so the offline picker still shows the full set of supported Codex slugs.

How to test

• pytest tests/hermes_cli/test_codex_models.py -q — passes (20 tests, including the new regression test).
• pytest tests/hermes_cli/test_codex_cli_model_picker.py tests/hermes_cli/test_model_validation.py tests/cli/test_cli_provider_resolution.py tests/cli/test_fast_command.py -q — passes (139 tests covering picker credential discovery, model normalization, /fast model handling).
• Manual: with no live token reachable, run hermes /model on the openai-codex provider — the picker no longer offers gpt-5.2-codex, gpt-5.1-codex-max, or gpt-5.1-codex-mini.

What platforms tested on

• macOS on darwin-arm64 (local)

Fixes #23097

<!-- autocontrib:worker-id=issue-new-23f4925b kind=pr-open -->

Changed files

• hermes_cli/codex_models.py (modified, +9/-3)
• tests/hermes_cli/test_codex_models.py (modified, +22/-0)

---

PR #23189: fix(codex): trust live ChatGPT picker catalog

• Repository: NousResearch/hermes-agent
• Author: LeonSGP43
• State: open | merged: False
• Link: https://github.com/NousResearch/hermes-agent/pull/23189

Description (problem / solution / changelog)

What does this PR do?

Stops the authenticated Codex model picker from re-inserting synthetic forward-compat models when the live ChatGPT account catalog is already available.

Related Issue

Fixes #23097

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• make hermes_cli/codex_models.py treat the live authenticated catalog as authoritative when API models are returned
• keep forward-compat synthesis only for fallback paths where no live catalog is available
• update tests/hermes_cli/test_codex_models.py to cover both the live-catalog path and the fallback path

How to Test

1. Run UV_PYTHON=3.11 /Library/Frameworks/Python.framework/Versions/3.10/bin/uv run --frozen pytest -q -o addopts='' tests/hermes_cli/test_codex_models.py
2. Run UV_PYTHON=3.11 /Library/Frameworks/Python.framework/Versions/3.10/bin/uv run --frozen ruff check hermes_cli/codex_models.py tests/hermes_cli/test_codex_models.py
3. Verify get_codex_model_ids(access_token=...) returns the live API list unchanged, while the no-live-catalog fallback still preserves forward-compatible defaults

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: macOS 15.x

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

Screenshots / Logs

• 20 passed in 0.90s
• ruff check passed for the touched Codex picker files

Changed files

• hermes_cli/codex_models.py (modified, +9/-3)
• tests/hermes_cli/test_codex_models.py (modified, +25/-10)