cavecrew

Delegate coding tasks to specialized subagents that return compressed outputs.

Updated Jun 4, 2026
One-click install
npx skills add https://github.com/achmf/KostaHub --skill cavecrew-achmf
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cavecrew
Source: https://github.com/achmf/KostaHub/tree/main/.agents/skills/cavecrew
Command: npx skills add https://github.com/achmf/KostaHub --skill cavecrew-achmf

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill addresses the issue of context window exhaustion during complex coding tasks by delegating work to specialized subagents that return highly compressed, token-efficient outputs.

Core Features & Use Cases

  • Context Preservation: Reduces tool-result overhead by ~60% compared to standard agent outputs, allowing for longer, more complex development sessions.
  • Specialized Roles: Provides targeted subagents for code investigation (cavecrew-investigator), surgical file editing (cavecrew-builder), and diff auditing (cavecrew-reviewer).
  • Use Case: When performing a multi-file refactor, use the investigator to locate symbols, the builder to apply changes, and the reviewer to audit the diff, all while keeping the main conversation context lean and focused.

Quick Start

Delegate the current task to the cavecrew investigator to locate all definitions of the specified function in the codebase.

Frequently Asked Questions about cavecrew

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

FAQPage Schema
How do I prevent context window exhaustion during large-scale code refactoring?

Prevent context window exhaustion during code refactoring by delegating tasks to specialized subagents that return compressed, token-efficient outputs, reducing tool-result overhead by approximately 60%.

How does a subagent framework help with code investigation and surgical file editing?

A subagent framework helps with code investigation and surgical file editing by assigning targeted roles to locate symbols, apply changes, and audit diffs, keeping the main conversation context lean.

What is the best way to optimize token usage for long-running development sessions?

The best way to optimize token usage for long-running development sessions is implementing structured output contracts that ensure subagents return compressed results, keeping the main context within limits.

Can I use compressed subagents for multi-file refactoring and diff review?

Yes, you can use compressed subagents for multi-file refactoring and diff review by deploying an investigator to locate definitions, a builder to apply changes, and a reviewer to audit the diff.

Does delegating coding tasks to subagents require any specific dependencies?

Delegating coding tasks to subagents does not require specific dependencies, as the framework operates independently to facilitate code investigation, file editing, and diff review within existing development projects.

When should I avoid using context optimization subagents for development tasks?

You should avoid using context optimization subagents for simple, short-lived development tasks where the overhead of delegating to specialized roles outweighs the benefits of preserving the main context window.