ck:scout

Locate files and summarize code relationships across large codebases.

1|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/quanganh208/cookmate --skill ck-scout-quanganh208
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ck:scout
Source: https://github.com/quanganh208/cookmate/tree/main/.opencode/skills/scout
Command: npx skills add https://github.com/quanganh208/cookmate --skill ck-scout-quanganh208

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill reduces the time and token cost of finding relevant files and gathering context across large, multi-directory codebases by running coordinated parallel subagents and returning a concise aggregated report.

Core Features & Use Cases

  • Parallel Scouting: Spawn multiple internal Explore or external Bash subagents to search distinct directory scopes in parallel and avoid overlap.
  • Flexible Tooling & Chunking: Use Gemini/OpenCode CLIs or internal Explore agents with line-count chunking strategies, 3-minute timeouts, and fallback logic when external tools are unavailable.
  • Task Coordination & Aggregation: Register scout tasks for observability, deduplicate and merge findings, and produce a single scout report for onboarding, debugging, or large refactors.

Quick Start

Run the scout targeting "auth" to locate and summarize all authentication-related files across the repository using internal Explore agents.

Frequently Asked Questions about ck:scout

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

FAQPage Schema
How do I locate files and summarize code relationships across a large multi-directory codebase?

Parallel codebase scouting locates files and summarizes relationships across large projects by spawning multiple Explore or Bash subagents to search distinct directories concurrently and aggregating the results into a single report.

What is the best way to gather context for feature onboarding in a large codebase?

Coordinated parallel subagents gather context for feature onboarding by registering scout tasks to search specific directory scopes in parallel, deduplicating findings, and returning a concise aggregated report.

Can I use external Bash agents for codebase scouting?

Yes, you can use external Bash agents for codebase scouting. The system supports spawning external Bash subagents with line-count chunking strategies, 3-minute timeouts, and fallback logic when external tools are unavailable.

How do I configure Gemini for parallel codebase scouting?

You can configure Gemini for parallel codebase scouting by adding optional Gemini configuration settings in the .opencode/.ck.json file to manage external tool selection and chunking strategies.

Does parallel codebase scouting work without external tooling dependencies?

Yes, parallel codebase scouting works without external tooling by using internal Explore agents as a fallback. It applies line-count chunking and 3-minute timeouts to manage searches when external tools are unavailable.

Why does my parallel codebase scouting report have overlapping or duplicated findings?

Parallel codebase scouting prevents overlapping findings by assigning distinct directory scopes to each subagent, then deduplicating and merging the per-agent reports into a single scout report.