[WRONG BRANCH] Align vision sidecar runtime default to gpt-5.6-luna

[luvs01]

Motivation

• The management API and GUI expose gpt-5.6-luna as the vision sidecar default while the runtime still fell back to gpt-5.4-mini, causing an operator-visible mismatch when visionSidecar.model was unset.

Description

• Change the OpenAI vision-sidecar runtime fallback from gpt-5.4-mini to gpt-5.6-luna by updating DEFAULT_VISION_MODEL in src/vision/index.ts.
• Add a focused regression test that verifies an unset OpenAI vision model resolves to gpt-5.6-luna in tests/claude-sidecar-override.test.ts.
• Update the sidecar architecture inventory in structure/04_transports-and-sidecars.md to document the corrected OpenAI vision default.

Testing

• Ran the focused test: node_modules/@oven/bun-linux-x64-baseline/bin/bun test tests/claude-sidecar-override.test.ts, which passed.
• Ran typecheck: node_modules/@oven/bun-linux-x64-baseline/bin/bun run typecheck (bun x tsc --noEmit), which passed.
• Attempted the full test suite with bun run test; it encountered unrelated environment-sensitive failures in tests/cli-help.test.ts and network-dependent discovery tests, so the full run was not used to gate this focused fix.

---

Codex Task

Summary by CodeRabbit

• New Features

  • Updated the default OpenAI vision model to gpt-5.6-luna for improved vision processing.

• Documentation

  • Updated the documented default vision sidecar model to reflect the new default.

• Tests

  • Added coverage confirming vision sidecars use the updated default when no model is specified.

[github-actions]

✅ Deterministic PR hygiene checks passed.

[github-actions]

⏳ DRAFT

• wrong target branch (main); retarget to dev. UI screenshot required.

What to do

• Retarget this PR to dev — all contributions go to dev.
• Add a screenshot of the UI change to the PR description.

Its title has been prefixed with [WRONG BRANCH].
This pull request was already a draft. Its draft status will be preserved after every issue above is resolved.

[coderabbitai]

📝 Walkthrough

Walkthrough

The OpenAI vision default model changes from gpt-5.4-mini to gpt-5.6-luna. Documentation reflects the change, and a regression test verifies selection when no model is configured.

Changes

Vision default model

Layer / File(s)	Summary
Update and validate the vision default src/vision/index.ts, structure/04_transports-and-sidecars.md, tests/claude-sidecar-override.test.ts	The runtime and documentation defaults change to gpt-5.6-luna. A regression test verifies that an OpenAI vision sidecar without an explicit model selects this default during planning.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Suggested reviewers: lidge-jun, wibias

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the main change: aligning the vision sidecar runtime default with gpt-5.6-luna.

✨ Finishing Touches 💡 1

🛠️ Fix failing CI checks 💡

• Create stacked PR
• Commit on current branch

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/propose-fix-for-vision-sidecar-default

---

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5d2f0f4d07

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".