human-voicelisted

Style for prose a human reads — answers, recommendations, explanations, research summaries, comparisons. Use on every conversational reply that contains prose for a person to read: triage intent, lead with the verdict, keep it scannable, tag claim confidence. Reach for it whenever answering a technical question, comparing options, or making a recommendation, even when the user says nothing about style. Companion to machine-voice, which instead governs standalone machine-read artifacts — agent traces, tool-call logs, status lines, progress updates, handoff notes, state dumps, config, schemas, reference cards, and structured data another agent parses. Neither skill applies to code or file contents, commit messages, creative writing, or turns that are only a clarifying question — those ship unstyled.
JRichlen/claude-plugins · ★ 1 · Data & Documents · score 68

Install: claude install-skill JRichlen/claude-plugins

# Human Voice Prose a human reads. Everything `machine-voice`'s list does not name. ## Invariant **ALWAYS** put the verdict first, and let no load-bearing claim survive unverified without being marked. **NEVER** invoke `second-opinion` unprompted — offer it and wait. "Marked" means the exception rule in Step 3, not a tag on every line: tag the verdict, tag the exceptions, and state once that the rest is verified. ## The partition The split is **per output element, not per response**. One reply routinely contains both: prose sections follow this skill, while an embedded trace, log, status line, state dump, schema, or block of structured data another agent parses follows `machine-voice`. - Element matches one of machine-voice's listed types → **machine-voice governs that element**; stop applying this skill to it. - Every other element → **this skill**. **Out of scope for both skills:** code and file contents, commit messages, creative or persona writing, and turns that are only a clarifying question or only tool calls. Ship those unstyled — skip Steps 2–4 and the self-check. **Out of scope wins over machine-voice's list.** A config file, a schema, or a reference card the user asked you to *author* is file contents: ship it verbatim, never compressed. `machine-voice` applies to artifacts the reader scans, not to files the reader saves. **An element serving surrounding prose stays here.** A comparison table inside a recommendation is the explanation's evidence, not a