What problem does it solve? Manual pull request reviews are inconsistent and slow, especially when trivial changes receive the same scrutiny as complex ones. This Skill automates structured PR reviews on GitHub by scaling review depth to PR complexity, running deterministic checks, validating against linked specs, and posting findings as inline comments and formal reviews. ## Core Features & Use Cases - Tiered Review Depth: Auto-detects Quick, Standard, or Deep tiers from PR size, labels, and file paths, with manual override flags. - Deterministic Checks: Runs in-context checks for missing tests, unbalanced Markdown backticks, and broken path references. - Spec Validation: Detects spec links in linked issues and validates design conformance, constraint compliance, and scope boundaries. - Multi-Agent Deep Review: Dispatches the code-review plugin and a feature-dev security reviewer in parallel for complex PRs. - Use Case: A reviewer runs the skill on PR #42; it auto-detects the Deep tier due to a security label, runs deterministic checks plus a security-focused agent, posts inline comments, and submits a REQUEST_CHANGES review with structured findings. ## Quick Start Ask the AI to review pull request number 42 using the forge-review skill and let it auto-detect the appropriate review tier.