[AnnatarHe]

Problem

The Claude Code Review job is green, and the review finds real issues, but it leaves no comment on the PR. Run 41383496248 on #339 reported 4 findings and posted none. The findings appear only in the job log.

/code-review high <pr> --comment posts through mcp__github_inline_comment__create_inline_comment. With tool search on (Claude Code's default), that MCP tool is deferred: the agent sees only its name until it loads it with ToolSearch. When it skips that step, it treats the tool as unavailable and, as the skill allows, prints the findings instead of posting them. The same happened across all ShellTime repositories with this workflow.

Fix

• settings: '{"env": {"ENABLE_TOOL_SEARCH": "false"}}': the action writes this to ~/.claude/settings.json, so Claude Code loads MCP tools with their schemas from the start. I checked this with Claude Code 2.1.317 (the CI version) and a stub MCP server. With default settings the tool is listed as deferred and ToolSearch is present. With this setting it is directly callable and ToolSearch is gone.
• Assert the review actually ran: the step now fails when the result reports findings (N findings) but no comment was posted, either through the inline tool or a gh api …/comments / gh pr comment|review fallback. Against the past runs' logs, it passes the android run that posted 5 of 5 and fails the runs above.
• It drops the "launched no subagents" warning, which fired on every run checked: /code-review high reviews the diff from several angles inline (high effort → 8 inline angles), so a good review launches none.

Testing

This PR changes the workflow, so the action skips its own review here, because the workflow has to match the default branch. The fix shows on the first PR opened after this merges: inline comments from claude[bot], and comments posted: N in the assertion step.

🤖 Generated with Claude Code

---

^{Open in [code]smith · Need help on this PR? Tag @codesmith-bot with what you need.}

[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	⚠️ Failed 2026-11-03T14:59:23.020774Z We couldn’t complete this review. Comment @codex review to try again.	2605f60	PR opened

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.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Flag	Coverage Δ
unittests	86.56% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.
see 1 file with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[shelltime]

🤖 Claude Code · Review and merge changes across sibling repositories

Reviewed Claude-generated changes across sibling repositories, merged applicable work, and returned repositories to updated main branches.

💰 Cost	🔢 Tokens	⏱️ Duration	🧠 Models	💬 Prompts	✏️ Lines
$3.31	7.2M	43m (17m active)	claude-opus-5-5, claude-haiku-5-5	4	+74 / −0

_{7.2M tokens: 1.5K input · 40.8K output · 7.1M cache read · 133.3K cache write}

📝 AI summary

Inspected repositories under the parent directory, using the Android repository as the known-good reference, and checked whether Claude Code review changes in the other repositories were valid or required fixes. Bash commands handled repository inspection, validation, Git operations, and merging across the repositories.

Merged the reviewed changes, then switched repositories back to their main branches and pulled the latest remote updates. The only recorded file edit was a temporary helper script at scratchpad/patch_review.py, used to support the review workflow rather than modify a project source file.

View this session on ShellTime →

_{Updated 2026-11-03 15:30 UTC by ShellTime · totals for the whole session so far · the session page is visible to its owner}