Skip to content

fairy123456789/voice-with-temperature

v1.0.0MIT

Edit written prose for clarity while preserving the writer's lived detail, uncertainty, rhythm, and emotional temperature.

GitHubOpen Website

Voice With Temperature — Preserve Human Writing Voice

Human Edge icon

Voice With Temperature edits prose for clarity without washing out the person who wrote it. Here “voice” means writing voice—lived detail, uncertainty, rhythm, specificity, and emotional temperature—not speech, audio, transcription, or voice synthesis.

Included Skill

  • voice-with-temperature: diagnoses generic AI smoothing, preserves meaning and texture, and produces a clearer draft plus a concise account of consequential edits.

Use it for essays, reflections, technical writing, speeches, bios, or posts where conventional polishing makes the writing fluent but less true.

Example prompts

  • “Edit this reflection for clarity, but keep the awkward detail and uncertainty that make it mine.”
  • “This technical post sounds AI-smoothed. Restore sentence rhythm and specific observations without adding claims.”
  • “Give me a light edit and a stronger edit, then explain which emotional choices changed.”

Install and test

From a clone of the repository:

claude plugin validate .
claude plugin marketplace add . --scope local
claude plugin install voice-with-temperature@human-edge-skills --scope local

For Cursor local development, copy this complete plugin directory to ~/.cursor/plugins/local/voice-with-temperature, reload the Cursor window, and confirm the Skill appears under Customize. Marketplace listing is not implied.

From the repository root:

python scripts/sync_plugin_packages.py --check
python scripts/validate_plugin_packages.py

Safety and authorship

The Skill should not invent memories, experiences, expertise, citations, or certainty. Review the result for factual drift and consent before publishing personal or third-party stories. It improves writing; it is not an impersonation or audio-generation tool.

See PRIVACY.md for the plugin's data-handling disclosure.

Source and license

The canonical source is skills/voice-with-temperature. The packaged copy is generated and hash-checked; edit the canonical Skill, then run the sync command from the repository root.

MIT licensed. See LICENSE.