Opt-in ultra-compressed OUTPUT style for OMK — port of JuliusBrussee/caveman (MIT, pin 0d95a81). Cuts output tokens ~65% (measured, range 22-87%) by speaking terse while keeping technical terms, code, errors, API names, CLI commands byte-exact. OUTPUT ONLY: 0% input/context/thinking reduction; adds ~1-1.5k input tokens per turn; net loss on terse Q&A or per-request/credit billing (Copilot etc). Six levels: lite / full (default) / ultra / wenyan-lite / wenyan-full / wenyan-ultra; switch via…
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# caveman (OMK opt-in port)
Ultra-compressed **output** style. Upstream: [JuliusBrussee/caveman](https://github.com/JuliusBrussee/caveman) @ `0d95a81` (MIT, 2026-07-03). See `LICENSE` and `SOURCE.md` in this directory.
Shrinks what the agent **says**, not what it knows. Brain still big. Mouth small. Output-only — does not compress input, context, files, or thinking tokens. Adds ~1–1.5k input tokens per turn. On terse Q&A or per-request/credit billing, net-negative. See Honest Numbers below.
---
Respond terse like smart caveman. All technical substance stay. Only fluff die.
## Persistence
ACTIVE EVERY RESPONSE. No revert after many turns. No filler drift. Still active if unsure. Off only: "stop caveman" / "normal mode".
Default: **full**. Switch: `/caveman lite|full|ultra|wenyan-lite|wenyan-full|wenyan-ultra`.
## Rules
…
The line at the top of this page pins the label git:20260707.8abbd1f, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
A 17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
pass: Frontmatter block present
pass: Frontmatter declares a name
pass: Frontmatter declares a description
pass: Size between 200 bytes and 200 KB (12383 bytes)
pass: No zero-width or bidi control characters
pass: No instruction hidden inside an HTML comment
pass: No link to an exfiltration or paste host
pass: No credential-shaped string
pass: No instruction to send local credentials anywhere
pass: No text hidden with inline styles
pass: No prompt-injection phrasing
pass: No curl or wget piped into a shell
pass: No recursive delete of root, home or parent
pass: No instruction to read or print local credentials
pass: No base64 blob over 200 characters
pass: No link to a raw IP address
pass: No script tag
Source
GitHub
dmae97/omk · 148 stars · license MIT · pushed 2026-10-23 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_cyw4wsi5w4ydri53
GET https://markdownregistry.com/api/v1/resolve?ref=dmae97/omk/caveman
GET https://markdownregistry.com/api/v1/blob/36d4410d96b225d0f7ef3045566339f6d7451c88cdeb629a4bb663cfe970cad1
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
dmae97/omk · .omk/skills/clone-website/SKILL.md · Reverse-engineer and clone one or more websites in one shot — extracts assets, CSS, and content section-by-section and…
dmae97/omk · .omk/skills/context7-mcp/SKILL.md · Fetch current library and framework documentation through the Context7 MCP server instead of answering from training…
dmae97/omk · .omk/skills/gitreverse/SKILL.md · Turn a public GitHub repository into a single synthetic user prompt that someone might paste into Cursor, Claude Code…
dmae97/omk · .omk/skills/omk-computeruse/SKILL.md · Route OMK computer-use and Neo tasks by verified tool and model capabilities, without a second orchestrator. Use for…
dmae97/omk · .omk/skills/omk-engine/SKILL.md · Project-local OMK operating profile. Use when configuring this repository's OMK workflow, selecting a small skill set…
dmae97/omk · .omk/skills/omk-neo/SKILL.md · Entry point for Neo computer-use work in this checkout. Route browser, website, research, code-review and MCP-setup…
dmae97/omk · .omk/skills/ponytail-audit/SKILL.md · Whole-repo audit for over-engineering. Like ponytail-review, but scans the entire codebase instead of a diff: a ranked…
dmae97/omk · .omk/skills/ponytail-debt/SKILL.md · Harvest every `ponytail:` comment in the codebase into a debt ledger, so the deliberate shortcuts and deferrals…
dmae97/omk · .omk/skills/ponytail-gain/SKILL.md · Show ponytail's measured impact as a compact scoreboard: less code, less cost, more speed, from the benchmark medians…
hoangnguyen0403/agent-skills-standard · .agents/skills/caveman/SKILL.md · Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical…
bug-ops/zeph · .zeph/skills/caveman/SKILL.md · Switch to ultra-compressed, telegraphic output. Use when the user asks for terse, brief, minimal, "caveman", or…
eltmon/overdeck · src/lib/caveman/skills/caveman/SKILL.md · Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical…
vodailocz/kilo-kit-mcp · skills/productivity/caveman/SKILL.md · Ultra-compressed communication mode. Cuts token usage ~75% by dropping filler, articles, and pleasantries while keeping…
