review

Cross-validate research outputs by delegating critiques to alternate AI agents via the FeedRadar CLI.

Updated May 10, 2026
One-click install
npx skills add https://github.com/ozzy-labs/feedradar --skill review-ozzy-labs
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review
Source: https://github.com/ozzy-labs/feedradar/tree/main/src/claude-skills/review
Command: npx skills add https://github.com/ozzy-labs/feedradar --skill review-ozzy-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Cross-review and validate research outputs by contrasting results from different agents to ensure accuracy and consistency.

Core Features & Use Cases

  • Delegates a review task to a different agent to provide independent critique.
  • Enforces frontmatter and payload conventions via the engine SKILL at .agents/skills/review/SKILL.md.
  • Supports host-agent mode with --emit-payload for manual review stages.

Quick Start

Use a single instruction to initiate a cross-agent review workflow, prompting the system to fetch an alternate critique and merge results.

Frequently Asked Questions about review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I cross-validate research outputs from different AI agents?

Cross-validate research outputs by delegating independent critiques to alternate AI agents via the FeedRadar CLI, contrasting results to ensure accuracy and consistency across multiple viewpoints on a single research item.

What is the best way to automate peer review between multiple agents?

Automate peer review by using the radar workflow to orchestrate cross-agent critiques and enforce status transitions, relying on the engine backend to manage the review stages and merge the results.

Do I need the radar CLI to perform a cross-agent research review?

Yes, the radar CLI is required to coordinate cross-agent research reviews, as it delegates critique tasks to alternate agents and enforces frontmatter and payload conventions through the engine backend.

Can I use host-agent mode to manually control review stages?

Yes, host-agent mode supports manual review stages by using the --emit-payload flag to output payloads, allowing you to control the workflow and enforce frontmatter conventions before merging results.

How do I enforce frontmatter and payload conventions during agent reviews?

Enforce frontmatter and payload conventions by relying on the engine SKILL located at .agents/skills/review/SKILL.md, which validates the structure and status transitions of delegated cross-agent review tasks.

Are there limitations to cross-validating conclusions with a single research item?

Cross-validation applies to verifying conclusions on a single research item at a time, orchestrating critiques from multiple agent viewpoints rather than batch processing multiple items simultaneously.