[saioai]

An empty vector-store search currently prints No results.. This makes the result specific and states that indexing was not checked.

What changes

• Successful empty searches print No matches returned. and Indexing state was not checked.
• Nonempty results keep attributes, scores, server order, source details, and every returned chunk.
• API errors remain errors. Search makes no additional status request and does not echo the query.
• Association lookup and live indexing acceptance remain incomplete.

Commands

No new commands or flags. The message applies to default and explicit readable output, including pipes.

openai vector-stores search --vector-store-id vs_demo --query hello
openai vector-stores search --vector-store-id vs_demo --query hello --format json # existing machine output
openai vector-stores search --vector-store-id vs_demo --query hello --format raw # full API page
openai vector-stores files retrieve --vector-store-id vs_demo --file-id file_demo # check actual file status

Explicit formats, extraction, --raw-output, and --max-items follow current main.
--format json returns one array, including [] for empty results. JSONL retains independent records.
--max-items 0 suppresses result records; explicit JSON still returns [].

Code

pkg/custom selects the vector-store message and passes it to the shared renderer in internal/output.
The renderer checks iterator errors before writing empty-result text. Other commands retain their default message.
Generated requests, transformers, dependencies, and main.go are unchanged.
Public tests cover request configuration, formats, expiration, chunking, attributes, batch controls, recovery, and separate attachment/store/source deletion.
The guide includes a bounded live acceptance recipe with owned-resource cleanup. It has not run against the live service.
The demo recorder keeps its checks active when Python optimization is enabled.

Tested

Candidate a3db91c9 incorporates main e963b2f5.
Public-command tests cover paginated lists, renaming, metadata, complete details, request configuration, and later-page failures.
The focused vector-store suite passes on macOS: 93 test/subtest pass events, with no skips.
Base URL integration tests pass: 24 additional test/subtest pass events.
Nested input recovery, focused race checks, module verification, and the CLI build pass.
Independent adversarial review found no supported regression.
CI passes at a3db91c9, including Linux tests and Windows compilation.
Platform help checks and CodeQL also pass.

Earlier 7c6d522 evidence includes 20 recorder checks and Bash syntax validation.
The recorder rejects invalid captures even with Python optimization enabled.
The unchanged live acceptance recipe passed three local success/failure scenarios at b2adb690.
No live API calls ran. Native Windows vector-store execution and graphical terminal acceptance remain unverified.
Association lookup and live indexing acceptance remain separate F15 requirements.

Live acceptance recipe.

Demo

Real CLI binaries with matching synthetic responses on macOS. Before: 5739ce0f. After: f4587145.
These are colorless terminal replays, not native graphical terminal acceptance.

Before:

After:

Recording recipe.

[github-actions]

Castiron custom code

Evaluated main: e963b2f5b9d493313061e09273c6ab62ece426cd.

Trusted main policy: e963b2f5b9d493313061e09273c6ab62ece426cd.

✅ No new custom-code files detected.

0 mixed files remain; 0 existing customizations changed.

Compared e963b2f5b9d4 → a3db91c9f50b. Generated baselines verified.

A changed generated baseline means this report cannot reliably identify which handwritten lines changed.

Inspect the custom-code diff

Download the exact patch produced by this run (requires repository access):

gh run download 39307929583 --repo openai/openai-cli \
  --name castiron-custom-code-39307929583-1 --dir /tmp/castiron-custom-code-39307929583-1
git apply --stat /tmp/castiron-custom-code-39307929583-1/custom-code.patch
cat /tmp/castiron-custom-code-39307929583-1/custom-code.patch

Or reproduce it from an SDK checkout containing the vendored reporter:

git fetch --no-tags origin e963b2f5b9d493313061e09273c6ab62ece426cd a3db91c9f50bb8030ccbcdfc5dcc74148124df4a
python3 scripts/castiron/custom_code_report.py report \
  --base e963b2f5b9d493313061e09273c6ab62ece426cd \
  --head a3db91c9f50bb8030ccbcdfc5dcc74148124df4a --fetch --require-head-hash --public \
  --out /tmp/castiron-custom-code-a3db91c9f50b
cat /tmp/castiron-custom-code-a3db91c9f50b/custom-code.patch

This is the current full custom patch for mixed files, not an attribution of only the handwritten lines changed by this PR.

Full report and patch

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-19T18:55:29.720050Z	a3db91c	New commits
🔒 Security Review	✅ Completed 2026-10-19T18:54:08.892610Z	a3db91c	New commits

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.