skills-marketplace/plugins/doc-help/skills/ghost-writer/SKILL.md

name

description

ghost-writer

Use this skill whenever the user wants to improve, rewrite, or humanize text in a document or message — especially to remove AI writing patterns and make it sound natural, personal, and authentic. Trigger on phrases like: "make this sound less AI", "humanize this", "rewrite this naturally", "remove AI signs", "polish my writing", "improve this document", "edit this to sound like me", "make this less robotic", "clean up this text", "ghostwrite this", "fix the AI tone", "improve writing style". Also trigger when the user pastes a document or block of text and asks for edits, rewrites, or style improvements — even if they don't explicitly mention AI. When in doubt, use this skill: it is better to load it and not need it than to miss a writing improvement opportunity.

Ghost writer skill

Improves writing quality and strips AI writing patterns, producing prose that sounds genuinely human - clear, specific, rhythmic, and authentic.

---

Before anything else: check for a style profile

Look for style-profile.md in the skill folder. If it exists and has content beyond the template comments, read it now.

Treat the profile as the user's personal style constraints. They override the defaults below:

• Tone and formality from the profile replace the "keep existing register" default
• Signature phrases in the profile are preserved even if they appear on the banned vocabulary list
• Avoid list in the profile is added to the banned vocabulary list
• Sample writing in the profile is used as a voice reference during rewriting
• Sentence style in the profile overrides the rhythm guidance in section G

If no profile exists, or the profile is blank, continue with the defaults for this rewrite - but flag it briefly: "(No style profile set up yet - using general defaults for this pass. I'll ask about setting one up at the end.)" Then proceed straight to Step 1. Do not interview the user before delivering value; the profile offer belongs at the end (see Step 6).

---

Step 1: Identify the input

If the user referenced a specific file, read it in full before doing anything else. If they pasted text directly, use that. If neither is clear, ask: "What text should I rewrite? You can paste it or give me a file path."

Do not skim. Read the whole thing.

---

Step 2: Establish the style target

Ask: "Any specific voice or style to aim for - e.g. formal report, direct technical writing, conversational, journalistic? Or should I just clean up the AI tells and keep the existing register?"

If the user says "just clean it up," keep the existing register and apply the rules below. Hold this answer before triaging - it affects which patterns are problems and which are intentional.

---

Step 3: Scan and triage

Run the quick scan protocol in this order:

1. Ctrl+F for banned vocabulary (section A below) - flag every hit
2. Read the first sentence of every paragraph - are they all the same grammatical shape?
3. Count em dashes - more than 3 per page is a flag
4. Find one factual claim - is it specific or vague? (abstraction trap, section E)
5. Find one transition word - does it appear more than twice per page?
6. Read 30 seconds aloud - does it have rhythm, or is it a metronome?

Classify findings:

• High priority (most detectable, most trust-damaging): abstraction trap, banned vocabulary, hedging, formulaic structure, "serves as" substitutions
• Medium priority: sentence rhythm, transition word overuse, Latinate bias, passive voice
• Low priority: em dash count, rule-of-three overuse, -ing openers

Exit path: If the scan finds only 1-2 low-priority issues, do not force a rewrite. Report what was found and ask: "This text is mostly clean - I found [X]. Do you want me to fix just those, or do a full pass anyway?"

If the user only wants a quick pass, fix high-priority items only and note what was skipped. For a deep rewrite, work through all categories and read references/ai-signs.md and references/good-writing.md for the extended checklist.

---

Step 4: Apply the rewriting rules

Work through all categories relevant to what triage found. Preserve intentional stylistic choices - if the author consistently uses a construction across the whole text, treat it as a voice marker, not a tell.

---

A. Banned vocabulary - replace or cut

Every instance must be replaced with a plain word or the sentence rewritten.

Remove or replace: testament, pivotal, underscore, watershed, profound, solidifies, stands as, serves as, crucial, vibrant, enduring, intricate, seamless, comprehensive, robust, dynamic, breathtaking, stunning, nestled, groundbreaking, renowned, boasts, bolstered, garner, foster, enhance, delve, tapestry, rich (as in "rich cultural heritage"), landscape (metaphorical), ecosystem (metaphorical), interplay, meticulous/meticulously, align with, highlighting, showcasing, emphasizing, realm, beacon, multifaceted, commendable

Replace "is/are" substitutes with the plain verb:

• "serves as" → "is"
• "stands as" → "is"
• "marks" (as synonym for "is") → "is"
• "represents" → "is" or "shows"
• "features" (meaning "has") → "has"
• "maintains" (meaning "has") → "has"
• "offers" (meaning "has") → "has"

Remove vague authority phrases entirely:

• "industry observers note", "experts argue", "some critics argue"
• "researchers suggest" (unless a specific researcher is named)
• "it has been suggested", "it is worth noting", "it is important to note"
• "scholarship describes", "studies show" (without citation)

If the underlying claim is worth keeping, rewrite it with a named source or direct statement. If it cannot be made specific, cut it.

---

B. Banned sentence patterns

"Not X but Y" / "Not only X but also Y" Artificial contrasts that sound insightful but restate the obvious. State the point directly.

Before: "Not just a technical tool, it is also a communication platform." After: "It is both a technical tool and a communication platform."

Trailing -ing phrases that restate the clause Delete any trailing participial phrase that repeats what the main clause already said.

Before: "The system improved reliability, reducing errors across deployments." After: "The system improved reliability."

The rule of three (triplet adjectives or vague phrases) Three-part lists of vague adjectives are an LLM fingerprint. Use one specific detail or none.

Before: "a dynamic, innovative, and forward-thinking organisation" After: "an organisation that ships in two-week cycles"

The "Despite / However" pivot formula Before: "Despite its many advantages, the system faces challenges, but continues to improve." After: State the specific challenge as a concrete fact, or cut the paragraph.

Transition word saturation Remove or replace: furthermore, moreover, additionally, consequently, thus, hence, in addition, it is worth noting, in contrast. Replace with direct connectives ("but," "so," "and") or restructure without them.

---

C. Structural patterns to remove

"In conclusion" / "As we have seen" / "In summary" openers Delete. Start the final paragraph with its actual content.

"In today's world" / "In today's [adjective] landscape" openers Delete the opener. Start with the substantive claim.

"As X continues to evolve" / "With the rise of X" openers Delete. Start with what actually happened.

Separate "Challenges and Future Prospects" or "Looking Ahead" sections Integrate challenges as concrete statements into the relevant sections. Cut speculation.

Meta-commentary and structural narration Delete sentences that describe the document rather than saying the thing:

• "In this article, we will explore…"
• "As we can see from the above…"
• "This section covers…" (as a standalone sentence)

---

D. Formatting tells to remove

Bullet points with bolded lead phrases The pattern "Term: explanation" throughout a document is an LLM fingerprint. Convert to prose. If a list is genuinely needed, use plain bullets without bolded labels.

Excessive bold mid-sentence Bold is for terms being defined or genuine emphasis only. Remove bold from any word bolded simply because the LLM deemed it important.

Excessive headers One header per short paragraph is a structural tell. Consolidate sections. A paragraph does not need a heading unless it begins a genuinely new section.

Em dashes used for decoration or aside Replace em dashes (—) with a hyphen-minus surrounded by spaces ( - ), a comma, or restructure the sentence.

Symmetrical bullet lists Every bullet the same length, every list the same count, is a tell. Vary: let some bullets be one sentence, others three. Cut a list to two items or expand to six.

---

E. The abstraction trap (highest priority)

Every generic claim must be replaced with a concrete fact, number, name, or example. If you cannot make it specific, remove it entirely.

Before: "The platform has helped many organisations improve their processes." After: "The platform reduced invoice processing time at Morrisons by 40%." (if that fact exists) Or: cut the sentence.

Apply to every sentence containing:

• quantities without numbers ("many," "several," "numerous," "various")
• superlatives without evidence ("the most," "the best," "leading")
• outcomes without measurement ("improved," "enhanced," "transformed")
• time without specifics ("recently," "in recent years," "over time")

---

F. Voice, clarity and authenticity

Latinate bias Replace long formal words with short plain ones:

• "utilize" → "use", "facilitate" → "help", "commence" → "start"
• "demonstrate" → "show", "approximately" → "about", "endeavor" → "try"
• "in order to" → "to", "due to the fact that" → "because"

Passive voice Say who did what. "Mistakes were made" → "The team made mistakes."

Hedging If the author means it, say it. Cut: arguably, potentially, it could be said, in many ways, one could suggest, somewhat.

Treadmill effect Cut restatements. If the same idea appears twice, keep the best version, delete the rest.

Perfect grammar, zero personality Add a fragment. Use a contraction. Let a short sentence stand alone. Like this.

Preserve the author's voice Do not over-sanitize. Retain domain vocabulary the author uses deliberately. Where the original has opinions, let them land - do not hedge what the author clearly meant to assert. One opinion should land in the final version without softening.

---

G. Sentence rhythm

LLMs produce sentences of similar length and structure. After all other fixes, check rhythm:

• Are consecutive sentences the same length? Break one up or combine two.
• Do multiple sentences start with subject + verb? Vary the opening.
• Does every paragraph have at least one short sentence (under 10 words)?
• Are all paragraphs roughly the same length? Use a one-sentence paragraph for a point that deserves to stand alone.
• Read aloud: if every sentence takes about the same time to speak, the rhythm is broken.

---

Step 5: Deliver the output

For short texts (under 300 words): Return the full rewritten version inline.

For long texts (300+ words): Save to outputs/ with -rewritten appended to the filename (e.g. report.md → outputs/report-rewritten.md). Return the rewritten text and the change summary.

Always provide a change summary. Lead with the dominant pattern found - not a generic list:

Dominant issue: [the single biggest AI tell in this text]
Changes made:
- [what was changed and why]
- [any claims cut because they could not be made specific]

---

Step 6: Offer one follow-up

Do not offer a menu. Choose the most relevant follow-up based on what triage found, and offer that one thing:

• No style profile exists yet → this takes priority over the other options below. Offer: "You don't have a style profile set up - want me to run a quick calibration (a few questions, optionally using a writing sample) so future rewrites automatically match your voice? Just say 'calibrate my writing style' whenever, or paste a sample now and I'll use it." Do not push further if declined - the defaults are working, this is just an offer.
• Vocabulary was the dominant issue → "I can explain the top 3 AI patterns I found in the original, if that would help you avoid them in future drafts."
• Structure was the dominant issue → "I can show you a before/after comparison of the paragraph that changed most, if you want to see the difference."
• Rhythm was the dominant issue → "I can do a second pass focused purely on sentence rhythm if you want to push it further."
• Abstraction was the dominant issue → "Several claims were cut because they couldn't be made specific. Do you have the numbers or examples to fill those gaps?"

---

Calibration trigger

If the user says "calibrate my writing style", "tune my style", "learn how I write", "update my style profile", "update my profile", "refresh my style profile", or similar - stop the normal flow and run references/calibration.md instead. Do not ask for text to rewrite.

This applies whether or not a profile already exists yet - calibration.md handles both the first-time interview and updating an existing profile (see its "Re-calibration" section).

---

Picking up style signal outside calibration

Not all style information arrives through the dedicated calibration flow. During a normal rewrite, the user may reveal something worth keeping permanently:

• Self-description: "I usually write for engineers," "I'm blunter than this in real life."
• A standing instruction framed as general, not one-off: "always cut adverbs," "I never use exclamation points," "from now on keep it this direct."
• An unprompted writing sample: the user pastes or references something they wrote and says "this is how I actually sound" or "match this," without invoking calibration.

Distinguish lasting preferences from one-off requests. "Make this one more formal" applies to the current text only - do not persist it. "I'm always more formal than this" or "from now on..." is a standing preference - it should end up in the profile.

If it's ambiguous which one you're hearing, ask: "Should I remember that for future rewrites too, or just apply it here?"

When something should persist:

1. Read the current style-profile.md (create it from the template if it doesn't exist yet).
2. Update only the relevant field(s) - Tone, Formality, Sentence style, Vocabulary (Signature phrases/Avoid), Audience, Opinions, or append to Source documents/Sample. Do not touch unrelated fields and do not run the full calibration question set for a single piece of feedback.
3. Confirm in one line: "Got it - saved to your style profile so future rewrites default to that." Do not make a bigger deal of it than the moment warrants.

This is different from full calibration: calibration is a deliberate 4-5 question interview that builds or replaces the whole profile; passive capture is a lightweight single-field update that happens inline during other work.

---

Reference files

File	When to read
style-profile.md	At the start of every invocation - user's personal style constraints
references/ai-signs.md	Before a deep rewrite - extended checklist of 27 AI writing patterns with examples
references/good-writing.md	Before a deep rewrite - principles of strong human prose (Orwell, Strunk & White, Hemingway)
references/calibration.md	When the user triggers the calibration flow

Read style-profile.md on every invocation. Read the others only for deep rewrites or calibration.

---

What this skill does NOT do

• It does not check factual accuracy. Content is preserved as written.
• It does not add information that was not in the original.
• It does not rewrite code blocks, technical commands, or quoted material.
• It does not change the language (British/American English, register) unless asked in Step 2.