⚠️ CRITICAL: This tool ONLY returns an ANALYSIS PROMPT for YOU (Claude), NOT a final review. OUTPUT: You receive comprehensive context (project type, diff, commits, structure) formatted as a detailed prompt with instructions. YOUR JOB: Analyze that context and generate the actual code review foll...
AI agents use review to create or update resources in Pr — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Pr environment.
An AI agent can call review faster than any human can review — one bad instruction and it creates or modifies resources in Pr by the hundred, each call as confident as the last.
Risk signalsBulk/mass operation — affects multiple targets
Attacks that exploit this kind of access
⚠️ CRITICAL: This tool ONLY returns an ANALYSIS PROMPT for YOU (Claude), NOT a final review. OUTPUT: You receive comprehensive context (project type, diff, commits, structure) formatted as a detailed prompt with instructions. YOUR JOB: Analyze that context and generate the actual code review following the exact format specified in the prompt. The prompt includes everything you need: project context, full diff, testing framework, architecture details. YOU must read it, analyze the code changes, and write a concise review (max 10-15 lines) with sections: Summary, Critical Issues, Key Suggestions, Decision (APPROVE/REQUEST_CHANGES). DO NOT return the prompt itself - generate YOUR review. Works for ANY language/framework because context is provided. Use when user asks to review code or when creating PR with review. It is categorised as a Write tool in the Pr MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Pr MCP server in PolicyLayer and add a rule for review: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Pr. Nothing to install.
review is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the review rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for review. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
review is provided by the Pr MCP server (valentin-harrang/pr-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.