caveman
Julius Brussee builds skills that help AI agents review code, implement user interfaces, and cut token costs. His most popular work includes compressed code review that returns one line per finding with location, problem and fix, and an interface guide covering accessibility, animation, typography and color systems. Much of his catalog centers on the Caveman toolset, which measures and reduces agent token usage through compressing memory files such as CLAUDE.md, reporting session statistics, and wiring repositories through a usage gateway.
Caveman
23 skills · 117,984 stars · Apache-2.0
Cavecrew When to delegate to `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit) or…
Caveman Terse caveman voice: answer first, fluff gone, every technical fact kept.
Caveman Commit Write a Conventional Commits message compressed to intent only.
Caveman Compress Compress a memory file such as CLAUDE.md or a todo list into caveman format to save input tokens,…
Caveman Discover Find and label every LLM workflow in the repository so Caveman Cloud groups spend by workflow ins…
Caveman Evidence Review Read-only review of Caveman Cloud evidence: cost, Cave Score, workflows, traces, latency, errors,…
Caveman Explore Read-only repository explorer for cold-start orientation, broad cross-file localization, or when…
Caveman Help Quick-reference card for the three caveman skills and their commands.
Caveman Learn Act on a Caveman learn report - review the ranked token sinks, apply cost-lowering fixes with per…
Caveman Manage Inspect Caveman Cloud's experiment lifecycle and block unsafe execution.
Caveman Optimize Turn a Caveman optimization observation into an operator-chosen candidate with a paired baseline…
Caveman Review Compressed code review - one line per finding with location, problem and fix.
Caveman Setup Wire a repository through the Caveman Cloud gateway so every LLM request is measured, with no beh…
Caveman Stats Show recorded output and cache-read token usage and mode attribution for the current Claude Code…
Caveman Stats Show recorded output and cache-read token usage and mode attribution for the current Claude Code…
Investigate First Diagnose ambiguous failures before editing.
Lean Build Build feature work with high overbuilding risk.
Megacave Caveman in Classical Chinese: 文言文 register, far fewer characters, technical terms verbatim.
Migration Implement reversible compatibility-safe transitions.
Safe Refactor Restructure code while preserving behavior.
Surgical Patch Fix bugs and small behavior changes at the narrowest responsible layer.
Ultracave Caveman at maximum compression: fragments, one word when one word is enough, each fact once.
Verify And Stop Prove existing work meets acceptance conditions without expanding scope.
Cavekit
9 skills · 1,160 stars · MIT
Backprop Bug → spec protocol.
Build Plan-then-execute implementation against SPEC.md.
Caveman Caveman encoding for SPEC.md and spec-adjacent writes.
Check Read-only drift detector.
Deepen Optional design-improvement pass for when you have spare usage to drain.
Grill Calibrated interrogation of a fuzzy idea before it becomes a spec.
Research Gather external knowledge the spec needs and distill it into §R — the durable
research log — so b…
Review Adversarial senior review of the spec before any code is written.
Spec Create, amend, or backprop bugs into SPEC.md at repo root.
Skills
8 skills · 175 stars · MIT
Caveman Ultra-compressed communication mode.
Context Canary Install a per-turn canary signal (e.g.
Deslopify De-slop pass for any text.
Grill Me Calibrated grilling session for stress-testing a plan, design, idea, or decision.
Interface Kit Authoritative guide for implementing stunning, accessible, performant UI.
Junior To Senior Adversarial senior-engineer review for agent-generated plans, designs, and architectures.
Last 20 Percent Finds and finishes the last 20% of a built solution — the experiential layer agents skip: what th…
Loop Factory Run a spec-driven agent loop where coding tasks live as markdown specs that move through inbox →…