From plainly
This skill should be used when writing or editing non-fiction prose for a human reader — essays, blog posts, documentation, READMEs, marketing copy, release notes — so the output avoids LLM "style smells" and follows a positive prose standard. It applies when drafting, rewriting, or polishing any such text.
How this skill is triggered — by the user, by Claude, or both
Slash command
/plainly:writing-clean-proseThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
When writing prose for a human reader, write *toward* the positive standard and *away* from the weighted tells — without over-correcting. An occasional em-dash, triad, or judicious nominalization is fine; the goal is good writing, not mechanical avoidance.
When writing prose for a human reader, write toward the positive standard and away from the weighted tells — without over-correcting. An occasional em-dash, triad, or judicious nominalization is fine; the goal is good writing, not mechanical avoidance.
Read these references (same files the auditor uses):
references/positive-standard.md — what good prose is (the goal).references/tells-catalog.md — the weighted patterns to avoid (the smells).references/rewrites.md — gold before/after transformations.Re-read and remove: any tricolon used as filler, any booster word ("transformative/game-changing") not backed by a concrete claim, any uniform-rhythm passage, any "In conclusion"-style wrap-up.
npx claudepluginhub mapika/plainly --plugin plainlyWhole-repo audit for over-engineering: finds dead code, unnecessary abstractions, stdlib-replaceable dependencies. Outputs ranked findings and net line/dep savings.