Conciseness

The body is lean and table-driven with terse prose, but the manual/off configuration is explained twice ("For opt-in activation... set defaultMode to manual" and again "Set \"off\" to suppress... Use \"manual\" when bare /caveman should activate"), which is a minor trimmable redundancy keeping it at anchor 4 rather than anchor 5.

Actionability

Guidance is fully executable: exact triggers per mode and skill, copy-paste-ready `export CAVEMAN_DEFAULT_MODE=ultracave` and JSON config snippets, an explicit resolution order ("env var > repo config > user config > caveman"), and legacy value handling — matching anchor 5.

Workflow Clarity

As a simple single-purpose skill, its one action is unambiguous: "Display this reference card when invoked. One-shot — do NOT change mode, write flag files, or persist anything." There are no destructive or batch operations, so the simple-skill exception applies and the action plus explicit prohibitions merit a 5.

Progressive Disclosure

Sections are well-organized (Modes, Skills, Deactivate, Language, Configure, More) with a clearly signaled one-level external link ("Full docs: https://github.com/JuliusBrussee/caveman") and no bundle files to reference. However, the body runs ~56 lines — just past the under-50 guideline — and the dense default-mode configuration could live in the external docs, fitting anchor 4 rather than 5.

Specificity

"Quick-reference card for the three caveman skills and their commands" names the domain and one concrete deliverable, but does not enumerate the skills or commands themselves, so coverage is not comprehensive (anchor 3) rather than listing several specific actions (anchor 4).

Completeness

Both "what" (a quick-reference card for the three caveman skills and their commands) and "when" (explicit "Trigger:" guidance) are present, so it clears the missing-trigger cap; however, the "when" covers only direct invocation with no situational trigger phrases, fitting anchor 4 rather than the concrete situational triggers of anchor 5.

Trigger Term Quality

"Trigger: /caveman-help or \"caveman help\"" provides a natural, relevant phrase, but both forms are the same invocation and common variations or synonyms (e.g., "caveman modes", "caveman commands") are missing, matching anchor 3 rather than the broader coverage of anchor 4.

Distinctiveness Conflict Risk

"caveman help" is a distinct trigger with a clear niche, but the shared "caveman" keyword across sibling caveman skills creates minor overlap risk with closely related skills, matching anchor 4 rather than the minimal conflict risk of anchor 5.