speclet-council

Coordinate parallel expert reviews of a Speclet Council draft and generate council artifacts.

Updated Jan 9, 2026
One-click install
npx skills add https://github.com/danielcastro-dev/speclet --skill speclet-council
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: speclet-council
Source: https://github.com/danielcastro-dev/speclet/tree/main/.opencode/skill/speclet-council
Command: npx skills add https://github.com/danielcastro-dev/speclet --skill speclet-council

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Speclet Council Skill orchestrates parallel multi-model reviews of an existing .speclet/draft.md and outputs structured audit artifacts to accelerate governance before converting to a spec.

Core Features & Use Cases

  • Validate .speclet/draft.md exists before any work
  • Invoke four reviewers in parallel with per-reviewer timeouts and retries
  • Collect critiques and generate a Review Status Header detailing succeeded vs. failed models
  • Synthesize feedback using Thematic Clustering and collapsible HTML details
  • Write .speclet/council-session.md and .speclet/council-summary.md
  • Write optional .speclet/draft.review.md (full Council Review)
  • Support --dry-run mode with mocked reviewer outputs

Quick Start

Run the speclet-council skill after you have a .speclet/draft.md to generate a council review.

Frequently Asked Questions about speclet-council

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a parallel multi-model review on a draft document?

To run a parallel multi-model review, you need an existing draft file and a tool that invokes multiple expert reviewers concurrently. This Skill coordinates parallel critiques with per-reviewer timeouts and generates structured audit artifacts like session and summary files.

What is a council review for software specification drafts?

A council review is a formal, multi-perspective assessment that critiques a draft before converting it to a spec. It synthesizes feedback from multiple models using thematic clustering to validate requirements and generate deterministic artifacts.

How do I generate structured audit artifacts from markdown drafts?

You can generate structured audit artifacts by running a formal review process on your markdown draft. This Skill collects critiques from parallel reviewers and writes deterministic files including council-session.md, council-summary.md, and draft.review.md.

Can I test a parallel review workflow without sending requests to all models?

Yes, you can test the parallel review workflow using a dry-run mode. This mode executes the coordination process with mocked reviewer outputs, allowing you to validate the artifact generation and synthesis logic without invoking actual models.

What happens if a reviewer model fails during a parallel draft assessment?

If a reviewer model fails during a parallel draft assessment, the process includes retries and generates a Review Status Header. This header details succeeded versus failed models, ensuring you know which perspectives were successfully collected.

Do I need a specific file format to start a multi-perspective draft validation?

Yes, you need an existing draft file specifically located at .speclet/draft.md to start multi-perspective validation. The workflow validates this file's existence before invoking any parallel reviewers or generating council artifacts.