Spoken Draft Cleanup
Turns a raw dictation transcript into a clean, editable draft while preserving the writer's voice and form exactly. MECHANICAL ONLY — removes fillers and false starts, restores sentence-boundary punctuation and paragraph breaks, and FLAGS likely transcription errors as questions rather than silently fixing them. It does NOT restructure, restyle, reword, or improve. Produces the cleaned draft plus a cleanup log of every change and every flag. This is the strict intake bridge from "speak out my article" to an editable draft, run before the advisory-edit passes. Use when cleaning a dictation/voic
Install
/plugin marketplace add lyndonkl/claudeThe repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.
npx skills add lyndonkl/claude --skill spoken-draft-cleanupgit clone --depth 1 https://github.com/lyndonkl/claude.git /tmp/claude
mkdir -p ~/.claude/skills
cp -r /tmp/claude/skills/spoken-draft-cleanup ~/.claude/skills/spoken-draft-cleanupPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Spoken-draft cleanup The writer dictates essays out loud. The transcript that comes back is the writer's thinking, verbatim — including the "um"s, the restarts, the missing commas, and the words the speech-to-text engine guessed wrong. This skill does exactly one job: make that transcript *readable as a draft* without changing what the writer said or how they said it. It is mechanical hygiene, not editing. This is the intake step in the vault's write loop: `transcripts/ -> drafts/` ([[conventions|Vault Conventions]] §2). It runs **before** any [[advisory-edit]] pass. The division of labor is strict — this skill produces a clean `draft` (status: cleaned) that still sounds 100% like the writer; advisory-edit then critiques that draft without touching it. Cleanup is allowed to alter the text (mechanically); advisory-edit is not. ## The one rule that overrides everything **Mechanical only. Preserve voice and form. Never restructure, restyle, or reword.** If a change would alter *what the writer is saying* or *how they characteristically say it*, you do not make it — you flag it. The writer's word choice, sentence rhythm, paragraph logic, idioms, hedges, jargon, and ordering are all sovereign and out of scope. You are a transcriptionist's assistant, not an editor. When in doubt about whether a fix is mechanical or editorial: it is editorial. Flag it. ## What cleanup MAY do vs MAY NOT do | MAY (mechanical) | MAY NOT (editorial — flag instead, or leave alone) | |---|---|
Details
- Skill name
spoken-draft-cleanup- Repository
- lyndonkl/claude
- Path
skills/spoken-draft-cleanup/SKILL.md- Repo updated
- 1 September 2026
FAQ
What does the Spoken Draft Cleanup skill do?
Turns a raw dictation transcript into a clean, editable draft while preserving the writer's voice and form exactly. MECHANICAL ONLY — removes fillers and false starts, restores sentence-boundary punctuation and paragraph breaks, and FLAGS likely transcription errors as questions rather than silently fixing them. It does NOT restructure, restyle, reword, or improve. Produces the cleaned draft plus a cleanup log of every change and every flag. This is the strict intake bridge from "speak out my article" to an editable draft, run before the advisory-edit passes. Use when cleaning a dictation/voic
How do I install the Spoken Draft Cleanup skill?
Claude Code plugin: /plugin marketplace add lyndonkl/claude. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Spoken Draft Cleanup popular?
It lives in lyndonkl/claude, which has 200 GitHub stars and 25 forks.
Related skills
Abstraction Concrete Examples
200lyndonkl/claude
Builds structured abstraction ladders that translate high-level principles into concrete, actionable examples across 3-5 levels. Bridges communication gaps, reveals hidden assumptions, and tests whether abstract ideas work in practice. Use when explaining concepts at different expertise levels, moving between abstract principles and concrete implementation, identifying edge cases by testing ideas against scenarios, designing layered documentation, decomposing complex problems into actionable steps, or bridging strategy-execution gaps.
Academic Letter Architect
200lyndonkl/claude
Guides the creation of evidence-based academic recommendation letters, reference letters, and award nominations that combine concrete examples, meaningful comparisons, and genuine enthusiasm. Use when writing recommendation letters for students, postdocs, or colleagues, or when user mentions recommendation letter, reference, nomination, letter of support, endorsement, or needs help with strong advocacy and comparative statements.
Adr Architecture
200lyndonkl/claude
Documents significant architectural and technical decisions with full context, alternatives considered, trade-offs analyzed, and consequences understood. Creates a decision trail that helps teams understand why decisions were made. Use when choosing between technology options, making infrastructure decisions, establishing standards, migrating systems, or when user mentions ADR, architecture decision, technical decision record, or decision documentation.
Strategist Voice
200lyndonkl/claude
Provides the house style for analyst-grade strategist writing — third-person register with sparing first-person, no em dashes, no "not X, not Y, not Z" negation cascades, numbered footnote citations rather than inline source parentheticals, specific opinion-signaling phrases, and topic-forward paragraph structure modeled on voice patterns observed in Damodaran's Musings on Markets and Thompson's Stratechery. Use when consolidating working notes into a finished long-form strategist or analyst report that must read as written by a senior human analyst rather than an AI assistant.
Reflect
100adriannoes/awesome-agentic-ai
Spawn three parallel review subagents over the active transcript, surface learnings, and route each to a concrete edit on an existing skill. Use when the user says reflect.
Voice Check
200lyndonkl/claude
Scans a substacker draft line-by-line against the canonical voice-profile.md don't-list and signature moves. Emits phrase-level flags with location, quoted phrase, violation type, voice-profile citation, and up-to-2 suggested rewrites per flag. Use as pass-2 skill (voice) after structural-review completes, when a draft reads competent but not in the writer's voice, or when the writer asks "does this sound like me?" Trigger keywords — voice check, delve, unpack, paradigm shift, sounds AI, does this sound like me, voice violation.
Browse the most popular Claude skills, all images, video & audio skills, or write your own with the skills generator.