Index of every page on this site: llms.txt
All skills
Run it yourself

Drafts, proofs, and audits prose in the author's voice, matched to the reader and destination. Covers social posts, articles, emails, docs, READMEs, commit messages, and feature announcements, with evidence rules that forbid invented facts, quotes, or experiences, and a draft only default that never publishes or sends. Use when drafting, rewriting, proofreading, or auditing social posts, articles, emails, docs, READMEs, or commit messages, or when the user says write:, proof this, or make this sound less like AI.

Use this Skill: https://skilld.dev/gh/waynesutton/builder-skills/write

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencesformats.md

≈1.3k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Format recipes and illustrative edits

Load this when drafting a specific format. Each recipe sets defaults, not hard limits. The destination and the user's own preferences win.

Format recipes

Short social post

Lead with a clear observation or announcement, add supporting detail, and include a next step only when useful.

  • One main idea per post. Do not stretch one thought into a thread.
  • Aim for the standard 280-character format unless longer copy is requested. Check destination-specific counting rules, especially for links, before claiming an exact count.
  • No hashtags or emojis by default; use them when requested.
  • Use destination-native link formatting and preserve supplied URLs and handles.
  • Avoid screenshots of the author's own quotes as a substitute for substance.

Professional social post

Use a meaningful opening, enough context, and a concrete insight or result. A next step is optional. Keep paragraphs mobile-readable and language plain and professional. Personal stories must be supplied or confirmed. Use full names rather than importing handles from another platform unless requested.

Blog or article

Use a topic-first lead and descriptive, sentence-case headings. As planning defaults, aim for a headline under about 70 characters, sections under about 300 words, and about five main sections or fewer. These are guidelines, not hard limits; depth and destination may justify more.

Use H2s for main sections and H3s only when they improve navigation. Source technical claims. Organize around real reader questions rather than a generic introduction-benefits-conclusion template.

README

Cover purpose, getting started, usage, configuration, and relevant API details. State requirements and constraints. Make the first successful action easy to find. Use verified commands and code where available; otherwise label examples untested and disclose assumptions. Do not invent APIs, configuration keys, or successful test results.

Commit message

Follow the repository's convention when supplied. Otherwise use type: imperative subject, with an optional body explaining why and important effects.

Common types: feat, fix, docs, style, refactor, test, chore.

Aim for a subject of 50 characters or fewer with no terminal period. Accuracy takes priority over forced brevity. Describe the actual change, not an imagined diff.

Feature announcement

Explain what changed, who it helps, the problem, how it works, and how to try it. Keep it concise. Only claim availability, performance, or compatibility when verified. Link to relevant examples or documentation when supplied or checked.

Task-oriented documentation

State the goal, prerequisites, ordered steps, expected result, useful example, and related material as needed. Place warnings before affected actions. Match actual interface or API terminology. Use tested examples where possible and label untested ones clearly.

Practical tip

Show one pattern or action, include code when relevant, and explain why or when to use it. State important limits. Do not inflate a small tip into a universal rule.

Email

Use a plain, specific subject, usually under 60 characters. Put the point first, give necessary context, and make one clear ask or next step when appropriate.

  • Skip stock well-wishing openers unless the author wants that convention.
  • Avoid default postscripts, fabricated urgency, and exclamation piles.
  • For outreach, give a truthful reason the recipient fits. Do not imply a nonexistent relationship.
  • For decisions, lead with the decision and its practical effect.
  • Offer wording alternatives for high-stakes messages when asked or useful. Routine email usually needs one draft.
  • Use a supplied sign-off or an explicit review placeholder such as [Your name]. Do not invent an identity.
  • Return a draft; do not send without separate authorization.

Illustrative edits

These examples are invented demonstrations, not claims about real people, products, or past corrections.

Topic before biography

Before: "As someone who has spent years organizing workshops, I wanted to share a reminder about preparation."

After: "Send the setup instructions before the workshop so participants can arrive ready."

Use the revision only if the brief supports it. Do not turn an edit into an invented policy.

Contact privacy

Before: A public draft includes a private email address copied from internal notes.

After: "Contact the team through the shared inbox."

Use that alternative only if the inbox exists and is an authorized public route. Otherwise omit the line or request an approved route.

Problem before inventory

Before: "The tool has templates, reminders, and shared notes."

After: "If meeting decisions get lost, keep each decision beside its owner and due date in shared notes."

The actual tool must support the proposed workflow. Do not invent capabilities for a more concrete sentence.

Minimum-edit proof

Before: "the hackathon is live go build something real people use"

After: "The hackathon is live. Go build something real people use."

Only capitalization and punctuation changed, not the claim or voice.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts11d3 checks · Risk SAFE
  • Gen Agent Trust Hub11d

    This skill is a writing and editing tool designed to draft, proofread, and audit prose while maintaining the user's voice and privacy. It includes strong instructions to avoid inventing facts, respect user boundaries, and protect sensitive information. No security risks or malicious patterns were detected.

  • Socket11d

    No alerts

  • Snyk11d

    Risk: LOW · No issues

Signed by skilld at 853b196. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 6 days ago.

Activeupdated 2 weeks ago

README badge

README badge for waynesutton/builder-skills/write