Skip to main content

Caveman

Skill Verified Active

Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) 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. Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer", "save context", "compressed agent output".

Purpose

To enable agents to perform complex code-related tasks more efficiently by intelligently delegating to specialized subagents and reducing token consumption, thereby extending the agent's context and capability over longer sessions.

Features

  • Delegates to specialized subagents for code investigation, building, and review.
  • Compresses subagent output to reduce main context token usage by ~75%.
  • Provides structured output contracts for subagent results.
  • Offers decision guidance on when to use cavecrew subagents versus vanilla Explore or main thread actions.

Use Cases

  • When needing to find specific code definitions or usages across a project.
  • When performing small, scoped edits to one or two files.
  • When reviewing code changes for bugs or issues.
  • To manage LLM context window limitations during long coding sessions.

Non-Goals

  • Providing general prose-based code analysis or architectural opinions.
  • Performing deep code reviews requiring subjective feedback.
  • Handling large refactors or changes across more than two files via the builder agent.
  • Replacing the main agent's core capabilities when context is not a constraint.

Workflow

  1. Main thread determines if delegation is needed based on task complexity and context.
  2. Selects appropriate cavecrew subagent (`investigator`, `builder`, or `reviewer`) based on the task.
  3. Subagent executes, producing compressed, structured output.
  4. Main thread injects compressed output back into context, preserving session length.

Installation

npx skills add juliusbrussee/caveman

Runs the Vercel skills CLI (skills.sh) via npx — needs Node.js locally and at least one installed skills-compatible agent (Claude Code, Cursor, Codex, …). Assumes the repo follows the agentskills.io format.

Quality Score

Verified
99 /100
Analyzed about 22 hours ago

Trust Signals

Last commit2 days ago
Stars59.5k
LicenseMIT
Status
View Source

Similar Extensions

Lean Ctx

100

Context Runtime for AI Agents — 59 MCP tools, 10 read modes, 95+ shell patterns, tree-sitter AST for 18 languages. Compresses LLM context by up to 99%. Use when reading files, running shell commands, searching code, or exploring directories. Auto-installs if not present.

Skill
yvgude

Ctx Doctor

100

Run context-mode diagnostics. Checks runtimes, hooks, FTS5, plugin registration, npm and marketplace versions. Trigger: /context-mode:ctx-doctor

Skill
mksglu

Metal

100

Extract the conceptual essence of a repository as skills, agents, and teams — the project's roles, procedures, and coordination patterns expressed as agentskills.io-standard definitions. Reads an arbitrary codebase and produces generalized definitions that capture WHAT the project does and WHO operates it, without replicating HOW it does it. Use when onboarding to a new codebase and wanting to understand its conceptual architecture, when bootstrapping an agentic system from an existing project, when studying a project's organizational DNA for cross-pollination, or when creating a skill/agent/team library inspired by a reference implementation.

Skill
pjt222

Pathfinder

100

Map a codebase into feature-grouped flowcharts, identify duplicated concerns across features, and propose a unified architecture. Use when asked to "find the ideal path," unify duplicated systems, or audit architecture before a refactor. Emits a proposed unified flowchart plus per-system /make-plan prompts.

Skill
thedotmack

Codacy Audit

100

Codacy Cloud workflow for this repository -- run Codacy's analyzers locally before `git push` (mirrors what Codacy CI runs), and fetch/cluster Codacy issues for any PR via the v3 API. Use when the user mentions Codacy, "codacy analysis", `codacy-analysis-cli`, "codacy issues on PR", "fix codacy CI", "codacy markdownlint findings", or any Codacy gate failing on a netdata-org PR. Ships scripts analyze-local.sh (docker/binary runner for codacy-analysis-cli) and pr-issues.sh (paginated v3 issue fetch + group-by tool/pattern/severity/file). Token-safe -- CODACY_TOKEN never reaches assistant-visible stdout. Read-only by design in the current SOW; write actions (mark FP, mark fixed) are deferred.

Skill
netdata

Domain Extract

100

Extract domain knowledge from existing project sources and generate domain rules. Also handles vault sync and domain listing.

Skill
luiseiman

© 2025 SkillRepo · Find the right skill, skip the noise.