Sunset prompt-only review references

[nehal-a2z]

Summary

• switch the Claude code-review agent doc to coderabbit review --agent
• switch the Claude review command doc to use --agent as the primary review path
• align the command severity/result guidance with the structured --agent output

Testing

• not run (docs/prompt changes only)

Summary by CodeRabbit

• Documentation
  • Updated code review workflow to agent-based analysis
  • Refined severity categories: Critical now includes security vulnerabilities and data loss risks
  • Replaced "Suggestions/Positive" with "Warning" and "Info" categories
  • Updated fix recommendation process with enhanced remediation details

[coderabbitai]

Caution

Review failed

Pull request was closed or merged during review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Central YAML (base), Organization UI (inherited)

Review profile: CHILL

Plan: Pro Plus

Run ID: dcb3d992-2672-40d6-af95-fb17c8219db1

📥 Commits

Reviewing files that changed from the base of the PR and between 521a49f and be30715.

📒 Files selected for processing (2)

• agents/code-reviewer.md
• commands/review.md

📜 Recent review details

🔇 Additional comments (2)

commands/review.md (1)

54-54: Good migration to --agent path and clearer severity mapping.

The updated command, severity taxonomy, and fix-offer wording are coherent and align with a structured agent-driven review flow.

Also applies to: 69-71, 73-73

agents/code-reviewer.md (1)

51-51: Workflow update is consistent and well-scoped.

Switching both review execution and interactive resolution to coderabbit review --agent is clear and consistent with the migration goal.

Also applies to: 65-65

---

📝 Walkthrough

Walkthrough

The workflow documentation is updated to shift from coderabbit review --plain to coderabbit review --agent as the primary code review command invocation. Severity category labels are refined, replacing "Suggestions/Positive" with distinct "Warning" and "Info" categories, and "Critical" is clarified to include security vulnerabilities and data loss risks.

Changes

Cohort / File(s)	Summary
Review Workflow Documentation agents/code-reviewer.md, commands/review.md	Shifted primary code review invocation from coderabbit review --plain to coderabbit review --agent. Refined severity grouping: "Critical" now explicitly includes security vulnerabilities and data loss; "Suggestions/Positive" replaced with "Warning" and "Info" categories. Updated fix-application workflows to prioritize --agent findings with actionable remediation details.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Poem

🐰 Whiskers twitch with glee,
The agent takes the lead, you see,
No plain reviews, just smart and keen,
With warnings, info, critical sheen!
Changes smooth as carrot dreams— ✨

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'Sunset prompt-only review references' clearly and concisely describes the main objective of the PR: removing/deprecating prompt-only review references in favor of the agent-based approach.
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.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch sunset_promptonly

✨ Simplify code

• Create PR with simplified code
• Commit simplified code in branch sunset_promptonly

---

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