CtrlK
BlogDocsLog inGet started
Tessl Logo

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".

76

Quality

95%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is ultracave in JuliusBrussee/caveman

SKILL.md
Quality
Evals
Security

Quality

Content

100%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body exemplifies the standard it enforces: maximal token efficiency with zero loss of technical substance. Concrete bad/good pairs, an explicit never-cut floor, break-the-rules conditions, and a pre-send checklist make the guidance fully actionable, and no bundle files are needed for a skill of this scope.

DimensionReasoningScore

Conciseness

The body is written in the skill's own target style ('Respond terse like smart caveman. All technical substance stay. Only fluff die.') with every line load-bearing — a Floor list, six one-line rules, break conditions, and a two-line checklist. There is no padding and no explanation of concepts Claude already knows, matching the 'lean and efficient; every token earns its place' anchor.

5 / 5

Actionability

Concrete Bad/Good example pairs cover the common response cases ('Inline object prop, new ref, re-render. `useMemo`.'; 'Yes. Null check first.'), the Floor section enumerates exactly what is never cut (code, commands, paths, negations, numbers, units), and the Pre-send check is an executable self-check. For an instruction-only skill this is copy-paste-ready guidance covering the common cases; rules 5–6 lack examples but are self-evident one-liners, so this fits the top anchor rather than the 'minor gaps' anchor below it.

5 / 5

Workflow Clarity

As a simple single-purpose skill the single action is unambiguous — respond in fragments subject to the Floor — and the 'Pre-send check' section ('Two readings? Full sentence. Negations present? Payload verbatim?') serves as an explicit validation checkpoint before every response. No destructive or batch operations exist that would cap the score.

5 / 5

Progressive Disclosure

The skill is roughly 55 lines, single-task, with well-organized sections (Persistence, Floor, Rules, When to break the rules, Pre-send check) and no bundle files referenced or needed — all content legitimately belongs in SKILL.md. This matches the simple-skill exception: well-organized sections with no external references required.

5 / 5

Total

20

/

20

Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is strong: it states concrete behavioral capabilities, gives explicit invocation and termination triggers, and occupies a distinct niche with minimal conflict risk. It could edge toward perfection by mentioning what is preserved (code, numbers, error strings) or brief synonyms like 'terse', but it already clearly communicates what the skill does and when to use it.

DimensionReasoningScore

Specificity

The description names concrete behaviors — 'fragments, one word when one word is enough, each fact once' — plus persistence semantics ('Stays on until "stop caveman" or "normal mode"'). It falls short of comprehensive because preservation rules and exception cases exist only in the body, but it clearly exceeds the 'minor gaps' threshold of 3.

4 / 5

Completeness

It explicitly answers both what ('fragments, one word when one word is enough, each fact once') and when ('Invoke only with /ultracave or /caveman ultra') with concrete trigger phrases, matching the top anchor. Neither element is missing or only implied.

5 / 5

Trigger Term Quality

It includes the exact natural phrases a user of this mode would say: 'caveman', '/ultracave', '/caveman ultra', 'stop caveman', 'normal mode'. A few natural synonyms (e.g. 'terse', 'brief', 'short responses') are missing, matching the 'good keyword coverage; a few natural terms missing' anchor.

4 / 5

Distinctiveness Conflict Risk

A narrow personal-response-mode niche with unique command-based triggers ('/ultracave', '/caveman ultra') and explicit stop phrases; no plausible overlap with other skills, matching the 'clear niche with distinct triggers' anchor.

5 / 5

Total

18

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
JuliusBrussee/caveman
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.