fix(pr-review-toolkit): stop collector review misattribution

[cblecker]

Summary

• In a review of Breakout bigquery ci data plugin openshift-eng/ai-helpers#786, the review collector (requested as haiku/low) received two coderabbitai[bot] reviews from get_reviews. It returned them as three records with author set to the reviewer's login: it skipped the login filter and split one review into two records. That switched on follow-up mode with HEAD as the reviewed commit, so the delta was empty and non-critical findings were demoted as "unchanged since your review."
• The collector now runs at effort: 'high' instead of 'low'.
• The get_reviews instructions now say to copy author verbatim from user.login and to return one record per review. They also require a re-check against the tool result that drops any review whose user.login isn't an exact match. Bots never match, and an empty reviews list is the normal result.
• Bumps pr-review-toolkit to 2.5.1 and updates the design note on collector effort.

This change reduces the chance of misattribution but adds no deterministic provenance check: JS still trusts the collector's author.

Test plan

• Workflow script parses (body wrapped in an async function, node --check)
• claude plugin validate ./pr-review-toolkit
• uvx skillsaw --strict
• markdownlint on DESIGN_NOTES.md

This PR was written in part with the assistance of generative AI.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository: cblecker/claude-plugins/.coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: f9593486-8411-4c50-a37b-1c7f21546b23

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Autopilot is currently an internal CodeRabbit preview.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused mitigation, documentation, and patch version bump are internally consistent.

Review effort: Balanced
Findings: None

What changed in this PR

Mitigates collector misattribution of submitted reviews during follow-up detection.

Changes:

• Strengthens review filtering and record-mapping instructions.
• Raises collector effort from low to high.
• Bumps the plugin patch version and updates design documentation.

File	Description
review-pr.js	Hardens collector behavior and increases effort.
DESIGN_NOTES.md	Documents the new collector effort.
plugin.json	Bumps version to 2.5.1.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused mitigation, documentation, and patch version bump are consistent and complete.

Review effort: Balanced
Findings: None