caveman

Ultra-compressed communication mode. Cuts token usage ~75% by dropping filler, articles, and pleasantries while keeping full technical accuracy. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman.

Quality

95%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

96%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.

An exemplary terse, fully actionable skill body whose form matches its content — every token earns its place and the guidance is copy-paste concrete. The one notable gap is that two of the three reference files are orphaned from SKILL.md, including when_caveman_backfires.md which would naturally be pointed to from the Auto-Clarity Exception section.

Suggestions

In the Auto-Clarity Exception section, link when_caveman_backfires.md (e.g., 'Failure modes: see [references/when_caveman_backfires.md]') so the backfire guidance is reachable from the entry point.

Mention compression_principles.md alongside companion_tooling.md in the Tooling section, since it is the reference paired with the compressor script.

Drop the trailing 'Version/Derived' footer from the body — it duplicates frontmatter metadata and costs tokens.

Dimension	Reasoning	Score
Conciseness	The body practices its own compression: specific banned-word lists ('just/really/basically/actually/simply'), a terse pattern template, and no explanation of concepts Claude already knows. Every section earns its place; the only overhead is the trailing version footer duplicating frontmatter.	5 / 5
Actionability	Fully executable guidance for an instruction-only skill: explicit drop lists, abbreviation vocabulary ('DB/auth/config/req/res/fn/impl'), a response pattern ('[thing] [action] [reason]. [next step].'), a Not/Yes contrast pair, and two worked Q&A examples. Claude knows exactly what to do after reading it.	5 / 5
Workflow Clarity	A simple single-purpose skill where the single action is unambiguous: activation ('ACTIVE EVERY RESPONSE once triggered'), exit conditions ('stop caveman' or 'normal mode'), and a scoped Auto-Clarity Exception with a clear drop/resume sequence and a concrete destructive-op example. The destructive-batch validation cap does not apply to a communication-style skill.	5 / 5
Progressive Disclosure	The body is a lean overview that clearly signals references/companion_tooling.md (verified to exist, one level deep). However, the other two bundle references — compression_principles.md and when_caveman_backfires.md — are never referenced from SKILL.md, making a third of the reference material discoverable only by directory listing. Good structure with a real, if bounded, navigation gap.	4 / 5
	Total	19 / 20 Passed

Description

91%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.

A strong description: third-person voice, a quantified 'what', and an explicit 'Use when' clause with natural quoted triggers. The only weaknesses are the generic 'be brief'/'less tokens' triggers that overlap with other brevity skills and a 'what' that omits a few of the compression rule categories detailed in the body.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions — 'Cuts token usage ~75%', 'dropping filler, articles, and pleasantries', 'keeping full technical accuracy' — with minor coverage gaps (body rule categories like abbreviations and causality arrows are not reflected). Not 5 because coverage is not comprehensive; not 3 because it goes well beyond 1-2 actions.	4 / 5
Completeness	Explicitly answers both: what ('Ultra-compressed communication mode. Cuts token usage ~75%... while keeping full technical accuracy') and when ('Use when user says "caveman mode", "talk like caveman"...') with concrete quoted trigger phrases. Matches the anchor-5 example structure directly.	5 / 5
Trigger Term Quality	'caveman mode', 'talk like caveman', 'use caveman', 'less tokens', 'be brief', and '/caveman' give comprehensive coverage of natural phrases users would say, including synonym variants and the slash-command form. Nothing a user would naturally say to invoke this is obviously missing.	5 / 5
Distinctiveness Conflict Risk	The 'caveman' triggers form a clear, distinct niche, but 'be brief' and 'less tokens' are generic phrases that any brevity or conciseness skill would also claim — minor overlap risk with closely related skills. Not 5 because the conflict risk is more than minimal.	4 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

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: alirezarezvani/claude-skills
Path: engineering/caveman/skills/caveman/SKILL.md
Commit: 19392f7

Reviewed: 17 hours ago

Table of Contents

Discovery Implementation Validation

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.