Multi-model consultation for SDD specifications providing structured feedback. Coordinates parallel AI reviewers, synthesizes actionable insights, and categorizes findings by feedback type without modifying specs or executing fixes.
The `sdd-plan-review` skill provides a structured, multi-perspective review process for Spec-Driven Development (SDD) plans. It addresses the challenge of ensuring draft specifications are robust, complete, and risk-aware before implementation begins. By convening multiple AI reviewers, this skill identifies missing information, design concerns, potential risks, and other areas for improvement, producing a consolidated, actionable feedback report. It solves the problem of uncoordinated or siloed feedback, helping teams maintain alignment and clarity on spec quality before any coding or execution starts.
Key features of this skill include multi-model consultation, categorization of feedback by type, and synthesis of reviewer perspectives into prioritized insights. It captures review metadata and organizes findings into structured Markdown or JSON reports for downstream consumption. Importantly, the skill provides advisory recommendations on which subsequent skills should handle each finding, without modifying specifications, making approval decisions, or executing changes. It ensures long-running operations are managed efficiently, avoiding excessive polling or background execution that could degrade the user experience.
This skill is particularly useful for development teams practicing SDD, AI-assisted review processes, or anyone responsible for ensuring the quality and completeness of software specifications. Target users include architects, product managers, and engineers who need thorough, categorized feedback on draft plans. It is applicable in workflows where multiple perspectives are valuable and when structured, actionable insights are needed to guide subsequent updates, validations, or downstream tasks, ensuring the team makes informed decisions before implementation begins.
You start by providing a draft specification. The skill will convene multiple AI reviewers to analyze the spec and generate a structured feedback report.
No. The skill is advisory-only and does not edit specifications, update metadata, or apply fixes.
Feedback is captured in both Markdown and JSON formats, including categorized findings and metadata for downstream processing.
Yes. Some reviews may take up to 5 minutes to complete. All commands should be executed in the foreground with appropriate timeout settings to avoid unnecessary polling or spam.
It is intended for software architects, product managers, and engineers who need multi-perspective, structured feedback on draft specifications before implementation.
Quick Setup:
.claude/skills/Repository
tylerburleigh/claude-sdd-toolkit