cavecrew

Decide when to delegate tasks to caveman-style subagents for software engineering.

36|7|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/Mosaic-agent/Mosaic-fund-agent --skill cavecrew-mosaic-agent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cavecrew
Source: https://github.com/Mosaic-agent/Mosaic-fund-agent/tree/main/.agents/skills/cavecrew
Command: npx skills add https://github.com/Mosaic-agent/Mosaic-fund-agent --skill cavecrew-mosaic-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn cavecrew-investigator, cavecrew-builder, or cavecrew-reviewer instead of doing the work inline or using vanilla Explore. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller — main context lasts longer across long sessions.

Core Features & Use Cases

  • Spawn and coordinate three subagents: cavecrew-investigator, cavecrew-builder, cavecrew-reviewer.
  • Output contracts ensure deterministic results and easy auditing.
  • Use cases include locating code, surgical edits, and diff reviews with minimized context.

Quick Start

Ask cavecrew to delegate to a subagent to initialize a caveman-guided workflow.

Frequently Asked Questions about cavecrew

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

FAQPage Schema
How do I delegate tasks to subagents for AI agent orchestration?

AI agent orchestration delegates tasks to subagents by automating the decision of when to spawn investigator, builder, or reviewer subagents rather than executing work inline. This enforces a three-party output contract for predictable, auditable results.

What is the best way to reduce context usage during long coding sessions?

Delegating work to caveman-style subagents reduces context usage by returning caveman-compressed output that is roughly 60% smaller. Injecting this smaller tool-result back into the main thread ensures your main context lasts longer across long software engineering sessions.

When should I spawn subagents instead of doing inline code edits?

You should spawn subagents instead of doing inline code edits for tasks requiring efficient delegation, such as locating definitions, performing surgical edits, and executing diff reviews. The skill automates this decision-making process to ensure safe main-thread integration.

Does delegating work to subagents enforce output contracts for code reviews?

Delegating work to subagents enforces deterministic three-party output contracts for code reviews and other software engineering tasks. These contracts ensure easily auditable results and straightforward chaining rules for safe main-thread integration.

Can I use subagent delegation for locating code definitions and surgical edits?

You can use subagent delegation for locating code definitions and surgical edits by spawning a cavecrew-investigator or cavecrew-builder. This minimizes context consumption while maintaining predictable, auditable results through enforced output contracts.