fix(agent): forward reasoning_config to background review AIAgent fork

[nftpoetrist]

What does this PR do?

`_spawn_background_review()` in `run_agent.py` (line ~3611) inherits the parent session's `provider`, `model`, `base_url`, `api_key`, and `api_mode` when forking a review agent. `reasoning_config` was not forwarded.

On Codex Responses routes the missing field causes the forked review agent to fall back to the transport default (`medium` effort), so a session configured for `agent.reasoning_effort: xhigh` still generates medium-effort background review requests — wasting budget and ignoring the user's explicit preference.

One-line fix in `_spawn_background_review()` — passes `reasoning_config=self.reasoning_config`. When `reasoning_config` is `None` the behaviour is identical to before: harmless no-op for sessions without a reasoning override. Symmetric with the `api_mode`, `base_url`, and `api_key` fields that #16006 and #15884 already propagate on the same fork path.

Related Issue

Fixes #18871

Type of Change

• 🐛 Bug fix
• ✨ New feature
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests
• ♻️ Refactor
• 🎯 New skill

Changes Made

• `run_agent.py`: add `reasoning_config=self.reasoning_config` to `AIAgent(...)` in `_spawn_background_review()` (+1 line)
• `tests/run_agent/test_background_review.py`: add `reasoning_config = None` to `_bare_agent()` helper; add `test_background_review_inherits_reasoning_config` parametrized over `None`, `{"effort": "xhigh"}`, and `{"enabled": False}` (+49 lines)

How to Test

```bash
python3.11 -m pytest tests/run_agent/test_background_review.py -v --override-ini="addopts="
```

Checklist

Code

• Contributing Guide read
• Conventional Commits
• No duplicate PR
• Single logical change only
• pytest passing
• Tests added
• Platform: macOS

Documentation & Housekeeping

• Docs updated — N/A
• cli-config.yaml.example — N/A
• CONTRIBUTING.md/AGENTS.md — N/A
• Cross-platform impact — N/A
• Tool descriptions — N/A

[teknium1]

Thanks for identifying the missing forked-agent runtime field. The defect remains on current main, but the implementation location moved after this PR was opened.

Problems

• run_agent.py:1598-1625 is now only a wrapper; the live review-fork construction is agent/background_review.py:683-697, where reasoning_config is still absent. The PR's added line therefore needs to be relocated to affect current behavior.
• This matters on the reported route: agent/transports/codex.py:155-163 defaults missing reasoning_config to medium and only applies a supplied effort value.

Suggested changes

• Salvage the constructor argument into agent/background_review.py:683-697 and port the regression test to capture that current constructor call for None, enabled xhigh, and disabled configurations.

Automated hermes-sweeper review.

[teknium1]

Current main moved the review-fork constructor to agent/background_review.py:683-697 in 1f6eb1738c206e95c3e0641c3d8000a4d0be841b; run_agent.py:1598-1625 is now a wrapper. Relocate this argument to the extracted constructor so the fix reaches the live path.

[kshitijk4poor]

Thanks for the fix — this was the right change, and you had it early. The same fix has now landed on main via #64379, a salvage of #30532 (the competing PR that paired the identical one-line change with the Anthropic cache-parity framing, wire-level measurements, and a regression test in the existing cache-parity suite). Your PR was submitted first among the competing fixes for #18871; #30532's version was selected on test coverage and framing, not timing.

Current main also gates the inheritance on the non-routed path (if not _routed: — when auxiliary.background_review routes the review to a different model, the fork uses provider defaults instead of the parent's effort vocabulary).

Closing as superseded by the merged fix. Thanks again for the contribution.