# Humanizing Prose

> Use when the task itself is to draft, rewrite, edit, or preserve human-facing prose for clarity, attribution, or proportion; do not select merely because other work ends with a response.

- Skill: `lawzava/humanizing-prose` (Agent Skill)
- Install (CLI): `npx skillmds@latest add lawzava/humanizing-prose`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lawzava/humanizing-prose/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: lawzava (https://skillmd.com/u/lawzava)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/lawzava/humanizing-prose

---


# Humanizing Prose

Apply this contract to plans, task briefs, commits, responses, reviews, PRs,
docs, release notes, and errors. Lead with the outcome. Keep the concrete
change, decision, evidence, limitation, or next action. Remove padding, sales
language, generic optimism, and session history that changes no decision.

Preserve identifiers, numbers, commands, caveats, uncertainty, decisions, and
all source facts. Never invent facts for rhythm, certainty, or completeness.
Do not rewrite quotations, code, or command names. Leave already-direct prose
alone.

## Machine-prose markers

Remove these before delivery; each one reads as generated text to the
operator. Do not use em dashes (U+2014) or spaced hyphens as em dashes; use a
comma, a period, a colon, or a new sentence. Apply this even when the
surrounding archive uses them. Remove triads of adjectives, "not X but Y"
framing, rhetorical questions, "it's worth noting", "in today's landscape",
summary headers that restate the body, and closing offers or recaps. Vary
sentence length instead of stacking parallel clauses. Keep one register
across the piece.

Replace vague attribution with a named source, direct observation, or explicit
uncertainty. Make an evaluative claim name the actor, mechanism, scope,
condition, or measurement that makes it specific. Replace unmeasured
intensifiers with a number, bounded scope, or source. Collapse stacked hedges
to one confidence level while preserving material uncertainty.

For an authorized review, PR comment, or issue update, publish only the current
verdict, material impact, and minimum evidence. Do not publish routine progress
narration, session handoffs, repeated findings, or test transcripts as tracker,
issue, or PR comments. Reply in the existing thread when possible.