scout

Analyze codebases locally to generate CLAUDE.md and decisions.sql artifacts.

Updated Apr 8, 2026
One-click install
npx skills add https://github.com/kelsi-andrewss/capstone-toolkit --skill scout-kelsi-andrewss
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: scout
Source: https://github.com/kelsi-andrewss/capstone-toolkit/tree/main/skills/scout
Command: npx skills add https://github.com/kelsi-andrewss/capstone-toolkit --skill scout-kelsi-andrewss

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Scout reveals what a codebase already requires before work begins by summarizing patterns, decisions, and constraints so planning stays aligned.

Core Features & Use Cases

  • Topic-driven inspection: Parse /scout arguments to gather clarifications, research artifacts, and bootstrap contexts.
  • Deep local analysis: Launch an Explore agent to map relevant files, dependency constraints, test patterns, and blast radius without any web access.
  • Bootstrap support: Initialize unfamiliar repositories to produce CLAUDE.md, decisions.sql, and refs that capture conventions.
  • Use Case: When onboarding to a new repo, /scout --bootstrap /path/to/repo generates the shared conventions Claude needs.

Quick Start

Ask Claude to run /scout <topic> with the necessary clarify or research files so it inspects the repo before implementation starts.

Frequently Asked Questions about scout

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

FAQPage Schema
How do I analyze an unfamiliar codebase to extract existing patterns and constraints before implementation?

To analyze an unfamiliar codebase and extract constraints, you can bootstrap the repository to generate a CLAUDE.md file and decisions.sql artifacts. This process maps dependency constraints and test patterns so your implementation planning stays aligned with existing conventions.

What is the best way to capture codebase decisions and conventions for consistent onboarding?

Capturing codebase decisions and conventions for onboarding involves generating shared reference files like CLAUDE.md and decisions.sql. These artifacts record testable assertions and dependency constraints, ensuring new implementations respect the repository's established architectural patterns.

Can I inspect a repository locally without web access to map its dependency constraints and blast radius?

Yes, you can inspect a repository locally without web access by running an Explore agent. The agent maps relevant files, dependency constraints, and blast radius entirely offline, ensuring topic parsing and decision-memory queries execute securely within your local environment.

How do I bootstrap a new repository to generate shared conventions for Claude?

You can bootstrap a new repository by running the bootstrap command with the target repository path. This initializes the unfamiliar repo, producing CLAUDE.md, decisions.sql, and reference files that capture the shared conventions Claude needs for consistent implementation.

Does codebase constraint analysis work with specific research artifacts and clarification files?

Yes, codebase constraint analysis works with research artifacts and clarification files by parsing them alongside the target topic. Supplying these artifacts when initiating the inspection ensures the Explore agent accurately maps relevant files and testable assertions for your specific implementation.

When should I not use a local codebase inspection approach for implementation planning?

You should avoid local codebase inspection when your implementation planning requires external web data or real-time API lookups, as this approach operates entirely offline. It is designed for mapping local repository constraints, not gathering external documentation during the analysis phase.