JuliusBrussee/caveman
🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman
Cut the chatter. Keep the work. A measured communication policy for coding agents: −32% output tokens, zero lost facts.
Appears on
Quick read
Latest capture 2026-08-25 03:05
6 paths
Agent instructions and tool configuration found in this repository.
Agent instructions
Agent workspace 3
Claude Code
Gemini CLI
6 observed captures since 2026-05-29. Observed captures are shown by default.
Stars from first capture +5
Observed captures only
All tracked data
Observed snapshots
Observed snapshots
Nearest indexed repositories by embedding similarity.
🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman
Make AI output sound human. Strips AI-isms (sycophancy, stock vocab, hedging stacks, em-dash pileups), preserves code/URLs/headings. Plugin for Claude Code, Cursor, Windsurf, Codex, Cline, Copilot, Gemini.
Ouroboros — self-creating AI agent. Born Feb 16, 2026.
MCP server that gets Claude to 97.9% (188/192) on a real coding benchmark at -80% active tokens and -83% wall time, vs 78.3% plain. Structural code navigation + persistent memory engine. Works with every MCP client.
Autonomous software engineering fleet of AI agents for production-grade PRs on AgentField: plan, code, test, and ship.
Context management for Claude Code. Hooks maintain state via ledgers and handoffs. MCP execution without context pollution. Agent orchestration with isolated context windows.