agentleFS
gleicongleicon/devskills/skills/ds-clarity-review/SKILL.md

ds-clarity-review

Skill20 starsChanged 48 days agoMITView on GitHub

Review prose for clarity against the Federal Plain Language Guidelines — any text: docs, comment prose, commit text, UI and error strings. Not `/ds-deslop` (code slop) or `/ds-comment-review` (whether a comment earns its place): this judges only whether a reader understands the words. Reports before/after findings; `--fix` applies the rewrites.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Report: No edits unless --fix.

    line 22
  2. Distinctive:

    Never count by eye — every number in a finding comes from the tooling.

    line 31

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

When invoked, review the prose in scope against the Federal Plain Language Guidelines, applied with Grice's maxims of manner and quantity in mind. The guidelines are the rule set — bring no private checklist, and order findings by reader impact (what most misleads first), not by which guideline fired. Technical terms the reader already owns count as plain language.

Boundaries: /ds-deslop removes code slop from the branch diff; /ds-comment-review judges whether comments earn their place; /ds-doc-quality-review audits documents as artifacts — accuracy against the code, links, coverage, bloat. This skill has one lens: does a reader understand the words? It optimizes for being understood, not for how the text sounds — reading-as-machine-written is /ds-humanize's lens — and it never judges code.

Arguments

  • Positional args are scope (files, directories, globs). With no scope, review the prose changed on the current branch.
  • --full → review all prose in the repo instead of the branch's changes. Explicit positional scope still wins.
  • --fix → apply the reported rewrites in place; exceptions stay untouched. After applying, confirm the diff touches prose only — code, identifiers, URLs, and quoted material must be byte-identical; revert anything else.

Process

  1. Collect the scope's prose. Fenced code blocks, inline code, URLs, and quoted material keep their exact form — never a finding, never edited.
  2. Run the drift pass mechanically. While reading, collect candidate pairs that may be one concept under two names ("calc type" / "calculation type"). grep the scope for each variant — the tool produces every count and site list. Then judge each enumerated pair: term drift, or genuinely two concepts.
  3. Review the prose against the guidelines, Grice's manner and quantity in mind — at sentence and document level both (the minor point given the most words is a quantity finding).
  4. Report. No edits unless --fix.

Output

For each finding: file:line, before (quoted), after (the rewrite). Where a rewrite would lose real nuance, don't rewrite — report it as an exception and say what would be lost. Ambiguity a reader would act on wrongly outranks structure, which outranks word choice. A short high-conviction list beats a long pedantic one.

Rules:

  • The guidelines govern; do not layer house rules on top of them.
  • Never count by eye — every number in a finding comes from the tooling.
  • Prose only. Code, identifiers, URLs, and quoted material stay byte-identical even under --fix.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.