Scribe

Make AI-assisted drafts sound like you wrote them.

Scribe is a Claude Code plugin for authors, essayists, and editors. It finds the patterns readers recognize as AI prose and rewrites them in a voice that fits your book.

/plugin marketplace add raymondware/kern
/plugin install scribe@raymondware
/scribe:revise

Paste a passage after /scribe:revise. Scribe returns the revised passage with before-and-after sameness scores on a 0-100 scale.

For a focused edit, use /scribe:humanize with a passage or file path. It removes stiffness while keeping your voice and meaning. For pasted text it returns just the edited passage; for a requested file edit it changes the prose in place. It skips the agent pipeline, pattern draw, audit log, and score, and it leaves prose that already works unchanged.

In today's fast-paced world, leadership isn't about authority; it's about influence. Great leaders don't just manage teams, but inspire them. By leveraging empathy, clarity, and vision, they unlock the full potential of their organizations. It's important to note that this is not a destination. It is a mindset.

Teams follow the manager who listens before deciding. She asks what is blocking the work, says plainly which goal matters this quarter, and then lets people do it.

18 agents · 31 anti-patterns · 5 writer personas · a different critique subset every run

AI-generated writing converges on the same cadence. "Delve into the tapestry of." "It's important to note that." "Not just X, but Y." "This is not X. It is Y." Em-dashes everywhere. Three-word lists. Conclusion paragraphs that restate the previous five. Hedging openers.

Scribe combines persona-specific writing rules with critic reviews of those habits.

Add the raymondware marketplace, then install Scribe:

/plugin marketplace add raymondware/kern
/plugin install scribe@raymondware

After you restart Claude Code, the /scribe:* commands are available.

# Generate a passage from scratch
/scribe:write opening chapter for a popular nonfiction book about how cities forget their own histories

# Audit existing prose
/scribe:audit
# (paste the manuscript section to audit)

# De-AI an existing draft
/scribe:differentiate
# (paste the draft)

# Run only the anti-pattern selector (verifies rotation)
/scribe:draw memoir essay about my year in Buenos Aires

Scribe picks a persona from your description, then loads that persona's rules for voice, rhythm, vocabulary, and structure. There are five personas:

Each pipeline run checks a different subset of anti-patterns. Scribe reads the audit log (state/draws.jsonl) and leaves out the previous run's subset, except for three negative-parallelism patterns that are included every time. So two runs on the same prompt are critiqued against different subsets. The focused /scribe:humanize skill doesn't use this rotation.

Rhythm, voice, cliche, and AI-ism critics run in parallel. A pacing critic joins for chapters and essays and always runs for /scribe:review. The synthesizer merges their output, removes duplicates, and computes a consensus sameness score.

Scored commands rate prose from 0 to 100 by how closely it resembles AI-default writing. Lower is better:

The score is a craft signal, not an AI-authorship detector.

Scribe ships with 31 anti-patterns across three files:

rhythm-specialist voice-specialist vocabulary-specialist structure-specialist dialogue-specialist

draft-implementer voice-implementer

rhythm-critic voice-critic cliche-detector ai-ism-critic pacing-critic

conductor critique-synthesizer (Opus), anti-pattern-selector (Sonnet)

research-scout external-judge cross-persona-audit

See CLAUDE.md for the contributor guide.

MIT. See LICENSE.