Rip-PR

Overview

Rip-PR is a pull-request review tool that treats AI agents as first-class reviewers with reviewable identities, launched October 12, 2026 (47 HN points) as the first major PR review tool to make agent-authored reviews indistinguishable from human-authored reviews in terms of auditability.

The core problem: as AI code review becomes ubiquitous (GitHub Copilot, CodeRabbit, Cursor Review all ship agent-based PR review), there's no way to distinguish an agent's review from a human's review in a way that an auditor or downstream maintainer can verify. Rip-PR changes that by giving every review an explicit identity — either human:<github-user> or agent:<agent-id> — and signing it cryptographically.

The result: reviews are attributable, auditable, and traceable back to the specific agent or person who authored them, with the same guarantees an auditor expects from signed commits.

Features

• Agent identity signing — every review is cryptographically signed with either a human's key or an agent's identity key
• Mixed human + agent review threads — human and agent comments interleave naturally in the same review thread
• Identity verification — GitHub verifies agent identity keys on first registration; subsequent reviews are verified on signature match
• OWASP Agentic A07 mitigation — agent privilege escalation detection at the review boundary
• MCP 2.2 server — Rip-PR exposes review operations as MCP tools
• Framework-agnostic reviewer plugins — reviewers can be Claude Code, Cursor, Copilot, or any custom agent
• Diff-aware review — reviewers see only the relevant diff hunks, not the whole file
• Test-result gating — reviews can be gated on CI test results; failing tests block review approval
• Compliance audit log — full audit trail of who reviewed what, when, with what authority

Installation

# GitHub App
# Add "Rip-PR" from the GitHub Marketplace

# Or via Helm for self-hosted
helm install rip-pr rip-pr/server -n rip-pr

# Or via Docker Compose
docker compose up rip-pr

MCP Configuration

{
  "mcpServers": {
    "rip-pr": {
      "command": "rip-pr-mcp",
      "args": ["--github-token", "${GITHUB_TOKEN}"],
      "env": {
        "RIP_PR_API_KEY": "***"
      }
    }
  }
}

Exposed tools: rip-pr:review, rip-pr:comment, rip-pr:request_changes, rip-pr:approve, rip-pr:list_reviewers.

Agent-Authored Review Example

from rip_pr import AgentReviewer
from agentguard import Runtime

# Register an agent identity
runtime = Runtime(config_path="./policies.yaml")
reviewer = AgentReviewer(
    runtime=runtime,
    agent_id="agent:security-reviewer-v3",
    reviewer_key="./keys/security-reviewer-v3.pem",
)

# Review a PR; identity-signed, OWASP-v2.1-compliant
review = await reviewer.review(
    pr_url="https://github.com/org/repo/pull/42",
    focus="security",
    max_comments=10,
)

Compliance Integration

Rip-PR audit logs are exportable in formats accepted by:

• AgentGuard — direct import of review-level audit logs
• AgentHub Enterprise — SOC 2 evidence upload
• SwarmAuth — agent identity verification

This makes Rip-PR usable as part of a fully audited agent-in-the-loop review pipeline.

Pros

• First PR review tool with real agent identity and cryptographic signing
• Framework-agnostic reviewer plugins
• Compliance-ready audit logs
• MCP 2.2 native
• OWASP Agentic A07 mitigation built in

Cons

• Newer tool with smaller ecosystem
• Setup is more involved than a simple GitHub App (key management)
• Requires agents to have signed identities

When to Use

• Production PR review with compliance requirements
• Mixed human + agent review teams
• Teams needing cryptographic attribution of reviews
• Regulated environments requiring audit trails

Resources

• Website: https://rip-pr.dev
• GitHub: https://github.com/rip-pr/server
• GitHub App: https://github.com/marketplace/rip-pr