Skip to content

Skillselionwith God's help

By phase

IdeaResearch & spot demand
ValidateScope & prototype
BuildFrontend, backend, MCP
ShipTest, review, deploy
LaunchDistribution, SEO, ASO
GrowAnalytics & retention
OperateMonitoring & uptime

By type

Skills
Marketplaces
MCP servers
Workflows

Skills Marketplaces MCP Learn Brief

Browse all tools →By phase

Idea Validate Build Ship Launch Grow Operate

Explore

Skills
Marketplaces
MCP
Learn
Brief

This week for builders

Five minutes, every Monday — the tools, releases and tactics for shipping solo.

Email address

unsubscribe anytime.

Skillselion

An independent, curated directory of AI-coding tools for solo builders - skills, MCP servers, marketplaces and workflows for Claude Code, Codex and Cursor, organized by the indie build journey from idea to operate.

ResourcesBrowse Skills Browse MCP Servers Browse Marketplaces Browse Workflows Best-of shortlists Glossary State of AI Skills 2026 Weekly Brief

CommunityAbout Learn Feedback Privacy Policy Terms Cookie Policy Accessibility Cookie settings

Last updated June 10, 2026 · © 2026 Skillselion. Skillselion is an independent project and is not affiliated with Anthropic, OpenAI, Cursor, Claude, Claude Code, or Codex.

Skills/SHAWNCHEE/CAVEMAN SKILL

shawnchee/caveman-skill

1 skill · 534 installs · 65 stars · GitHub

Install

npx skills add https://github.com/shawnchee/caveman-skill

Skills in this repo

1CavemanCaveman is a journey-wide agent skill that forces a token-saving communication mode: direct sentences, zero greetings, and silent tool use while still delivering full technical artifacts. Solo and indie builders install it when long-winded agent replies slow iteration, inflate context windows, or bury the actual diff or command in preamble. It fits every phase of the Prism journey—idea research, validate scoping, build and ship work, launch copy, growth experiments, and operate firefighting—because it changes how the agent talks, not what domain you are in. The SKILL.md is a compact rule set rather than a multi-step workflow: no checklists, gates, or chained follow-on skills. Expect shorter chat transcripts, faster scan of outcomes, and less accidental restatement of your question; it does not replace planning, security review, or structured skills like brainstorming when you need explicit design before code.534installs