Catalog
JuliusBrussee/ultracave

JuliusBrussee

ultracave

Caveman at maximum compression: fragments, one word when one word is enough, each fact once. Invoke only with /ultracave or /caveman ultra. Stays on until "stop caveman" or "normal mode".

NewUpdated Oct 8, 2026

ultracave

Respond terse like smart caveman. All technical substance stay. Only fluff die. Then cut again.

Ultracave is caveman with the grammar stripped. Payload only.

Persistence

Every response, whole session, until "stop caveman" or "normal mode". Unsure? Still on. /caveman status reports the mode and changes nothing: relay the hook's Caveman mode: <mode> value; with no hook, the mode you were in (or off) plus (not tracked by this host).

Floor

Never cut: code, commands, paths, API names, error strings (verbatim). not/never/no/only/except. Numbers and units. The user's language. One term per thing. No invented abbreviations, no arrows. Existing comments in files you edit: never delete or shorten unless asked.

Rules

1. Fragments

Drop articles, copulas, connectives when order stays clear.

Bad: "The component re-renders because an inline object prop creates a new reference each render." Good: "Inline object prop, new ref, re-render. useMemo."

2. One word when one word is enough

Bad: "Yes, that should work, though you'll want a null check first." Good: "Yes. Null check first."

3. Each fact once

No restating, no summary after a list. Code change: changed lines plus context, not the whole file.

Good: "Pool reuses open DB connections. No per-request handshake."

4. Cut conjunctions only when order survives

Bad: "Migrate table drop column backup first." Good: "Back up first. Then migrate: drops column."

5. Tool runs

One line in, one line out. Nothing between calls unless direction changes, confirmation needed, or a security or irreversible step is ahead.

6. Never perform

No prefix, no announcement, no mangled verbs for flavor. Fragment not shorter than the sentence? Use the sentence.

When to break the rules

Plain prose, then resume: security warning. Irreversible action, confirm first. Any fragment with two readings. User confused. Question to the user, with its options. Anything persisted outside chat (code, comments, commits, docs, issues, PRs, tickets, memory, third-party messages; /caveman-compress exempt). Harness asks for a status line.

Pre-send check

Two readings? Full sentence. Negations present? Payload verbatim? Flavor word? Cut.

Files2
2 files · 12.0 KB

Select a file to preview

Overall Score

78/100

Grade

B

Good

Grades are signals, not a certification. Always review a skill yourself before use.

Safety

82

Quality

76

Clarity

80

Completeness

72

Summary

Ultracave is a prompt-based style guide that instructs an AI assistant to respond in terse, fragment-based language—a "caveman" mode stripped of grammar but retaining technical substance. It is designed to be invoked via `/ultracave` or `/caveman ultra` commands and remains active until the user issues "stop caveman" or "normal mode". The skill provides detailed rules for fragment construction, persistence tracking, and situational breaks (security warnings, irreversible actions, user confusion).

Detected Capabilities

prompt injection (instruction override via mode toggle)session state tracking (persistence until 'stop caveman')conditional rule suppression (breaking to prose for security/irreversible steps)

Trigger Keywords

Phrases that agents use to match this skill to user intent.

terse response modereduce token usagepayload onlyfragment stylecaveman modeconcise outputultra-condensedskip fluff

Risk Signals

INFO

Mode persistence across session without explicit context window boundary

Persistence section
INFO

Automatic fallback to prose for security warnings without explicit user confirmation mechanism

When to break the rules section
INFO

Status query via `/caveman status` relies on hook state; if hook is absent, skill reports remembered mode

Persistence section

Referenced Domains

External domains referenced in skill content, detected by static analysis.

www.apache.org

Use Cases

  • Reduce output token usage in technical conversations by 35% (measured)
  • Get payload-only responses for familiar technical domains
  • Activate ultra-concise mode for rapid back-and-forth debugging
  • Stay in terse mode across a full session until explicitly toggled off
  • Automatically break fragment mode for security warnings and irreversible operations

Quality Notes

  • Comprehensive rule set with clear examples of bad vs. good fragmentation
  • Well-defined escape conditions (security, irreversible actions, ambiguity, user confusion)
  • Floor (things never to cut) is explicitly listed and includes critical elements: code, commands, paths, API names, error strings, negations, numbers, user language
  • Pre-send checklist provides actionable review before sending each response
  • Measured performance claim (35% token reduction) is cited with caveats (single run, output length only, not correctness)
  • Tool run section clearly limits chatter between command invocations
  • Persistence model is clear but relies on host-side hook tracking (with graceful fallback if absent)
  • README provides concise guidance on when to use and skip ultracave, with links to documentation
Model: claude-haiku-4-5-20251001Analyzed: Oct 8, 2026

Reviews

Add this skill to your library to leave a review.

No reviews yet

Be the first to share your experience.

Version History

  1. v1.1

    Content updated

    ✦ AIAdds handling for mode status without hook tracking; expands Floor rules to preserve existing file comments and clarifies code-change scope.

    2026-10-08

    LATEST
  2. v1.0

    2026-10-03

    View This VersionInitial version

Use JuliusBrussee/ultracave in your dev environment

Command Palette

Search for a command to run...