skills.rest
junechen7414
junechen7414COMMUNITY
Explore all skills by this author →
GitHub

cavecrew

Route coding tasks to investigator, builder, or reviewer subagents.

Updated Jul 2, 2025
Search Tags:#context management#subagents#delegation#code search#token efficiency#diff review
One-click install
npx skills add https://github.com/junechen7414/SpringBoot --skill cavecrew-junechen7414
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: cavecrew
Source: https://github.com/junechen7414/SpringBoot/tree/main/.agents/skills/cavecrew
Command: npx skills add https://github.com/junechen7414/SpringBoot --skill cavecrew-junechen7414

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you decide when to delegate work to compact caveman-style subagents instead of handling everything inline, reducing token usage and keeping long sessions from running out of context.

Core Features & Use Cases

  • Delegation Guidance: Chooses between investigator, builder, and reviewer subagents based on task scope.
  • Context Efficiency: Uses compressed outputs so subagent results consume less main-thread context.
  • Workflow Planning: Supports common patterns like locate → fix → verify for code changes and reviews.
  • Use Case: If you need to find where a symbol is defined, edit one or two files, or review a diff, this Skill helps you route the task to the right subagent quickly.

Quick Start

Ask the assistant whether to use cavecrew for a delegation-heavy coding task and let it recommend the right subagent path.

Frequently Asked Questions about cavecrew

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

FAQPage Schema
How do I reduce context token usage when handling long coding sessions?▼

Context management for long coding sessions is handled by delegating coding work to compressed subagents. This routes tasks like code search, narrow edits, and diff review to compact subagents, returning structured outputs that consume less main-thread context.

What is the best way to delegate code search and narrow edits to subagents?▼

Delegating code search and narrow edits is best handled by routing tasks to investigator or builder subagents. This approach applies structured output contracts and scope limits to ensure the subagent returns compressed results efficiently.

When should I use a reviewer subagent instead of inline diff review?▼

Using a reviewer subagent for diff review is ideal when verifying code changes without exhausting main-context tokens. It applies structured output contracts to evaluate the diff within defined scope limits, keeping the main thread uncluttered.

How does routing work between investigator, builder, and reviewer subagents?▼

Routing between investigator, builder, and reviewer subagents works by matching task scope to the correct workflow. Investigators handle code search, builders manage narrow edits, and reviewers verify diffs, all returning compressed outputs to maintain token efficiency.

Can I use compressed subagents for locate, fix, and verify code change workflows?▼

Compressed subagents support locate, fix, and verify code change workflows. This workflow planning delegates the locate step to an investigator, the fix to a builder, and the verify step to a reviewer to minimize token usage.

Related Skills

Explore More Software Engineering Skills→
rvdbreemenrvdbreemen
199

cavecrew

Route code tasks to investigator, builder, or reviewer subagents.

Community
Basic
artgaurav16420-ossartgaurav16420-oss

cavecrew

Delegate exploratory tasks to caveman-style subagents for compressed outputs.

Community
Intermediate
achmfachmf

cavecrew

Delegate coding tasks to specialized subagents that return compressed outputs.

Community
Intermediate
shirulotshirulot

cavecrew

Delegate coding tasks to specialized subagents with compressed output contracts.

Community
Intermediate