Caveman

Cuts spec documentation down to about 25% of original token count while keeping precision intact. Drops articles, filler words, and auxiliary verbs, then leans on symbols like → for "leads to" and ∴ for "therefore". Works well for SPEC.md files, invariants, and bug tracking tables where you're writing for yourself or your team, not external stakeholders. Leaves code, paths, error strings, and identifiers completely untouched. The compression is aggressive but reversible in your head, which makes it useful for high-churn design docs where you're iterating fast and token budgets matter. Switch back to prose when writing commits, PRs, or anything a non-team member will read.

skills.sh installs: 2.8K
GitHub stars: 1.2K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add juliusbrussee/cavekit --skill caveman --agent claude-code

Installs into .claude/skills of the current project.

Files

caveman — spec encoding

Applies to SPEC.md writes, spec-referencing prose, backprop entries. Does NOT apply to code, error strings, commit messages, PR descriptions.

GRAMMAR

Drop articles (a, an, the).
Drop filler (just, really, basically, simply, actually).
Drop aux verbs where fragment works (is, are, was, were, being).
Drop pleasantries.
No hedging (skip "might", "perhaps", "could be worth").
Fragments fine.
Short synonyms: fix > implement, big > extensive, run > execute.

SYMBOLS

Prefer over words:

→   leads to / becomes / on <x>
∴   therefore / fix
∀   for all / every
∃   exists / some
!   must / required
?   may / optional / unknown
⊥   never / forbidden / nil
≠   not equal
∈   in
∉   not in
≤   at most
≥   at least
&   and
|   or
§   section reference