mimus vs the alternatives

vs linters

	linters	mimus
checks	code shape	claims about reality
docs	mostly ignored	first-class source
output	style/correctness	claim + evidence + fix
rules	per-language config	rules/*.toml, evidence-bound

They complement each other — a linter says the code is shaped right; mimus says the docs tell the truth.

vs LLM doc review

	"review the docs" prompt	mimus
coverage	broad, vague	only what's checkable
false positives	high	evidence-or-silence
CI gate	unsafe	mechanical tier is safe
cost	per-run	mechanical tier is free

The LLM still has a role — as the check-llm adapter for the semantic tier, where findings are advisory by design.

vs doctest-style tools

	doctests/link-checkers	mimus
scope	code blocks only	all factual prose
claim types	one or two	path/cmd/symbol/count/link
memory audit	n/a	.kuda/ card trust scores