grimoire-pr-review

Automated pull request reviews apply a multi-persona framework—product, engineer, security, QA, and data—to the actual diff, generating structured findings ready for inline PR comments. The review fetches PR metadata, computes the real diff, and grounds assessments in linked artifacts.

Multi-persona code review analyzes a pull request diff across distinct roles like product, engineer, security, QA, and data. It produces a formal, structured report with evidence from linked materials, formatted specifically for posting as PR comments.

Provide a PR URL or number to fetch the PR and compute the real diff. Load linked grimoire artifacts via the Change trailer to ground the assessment in actual changes, then output structured findings across multiple personas for PR comments.

Yes, AI-assisted PR review includes security and QA personas in its multi-persona framework. It computes the real diff and applies these specific lenses to generate structured findings, ensuring the assessment covers security vulnerabilities and quality assurance concerns.

Multi-persona PR review operates with no external dependencies. It fetches PR metadata, computes the diff, and loads linked grimoire artifacts via the Change trailer internally to produce structured findings.

Grounding code review findings in actual changes requires computing the real diff and loading linked grimoire artifacts via the Change trailer. This approach ensures the multi-persona assessment is based on the actual modifications rather than just the PR description.