Harvey Specter Writing Style

Rewrite or draft text in a confident, concise, and strategically assertive style inspired by Harvey Specter, ideal for professional negotiations and sharp co...

Security Scan

The skill is an instruction-only writing-style transformer that is internally consistent with no unusual permissions, installs, or data-access requests.

Purpose & Capability

Name/description match the instructions: the SKILL.md defines rewriting/drafting in a Harvey Specter–inspired voice and requires no external credentials, binaries, or unrelated resources.

Instruction Scope

Runtime instructions are narrowly scoped to rewriting/drafting text, include explicit guardrails (avoid verbatim Suits quotes, keep professional, convert intimidation to firm boundaries), and do not instruct reading system files, environment variables, or calling external endpoints.

Install Mechanism

No install spec or code files; instruction-only skills have minimal on-disk footprint and no downloads or package installs are requested.

Credentials

The skill requests no environment variables, credentials, or config paths — consistent with its stated purpose of text transformation.

Persistence & Privilege

always is false and there is no request to modify other skills or system settings; default autonomous invocation is allowed but not exceptional here.

Assessment

This skill appears safe from a permission and install perspective. Consider these practical points before enabling: (1) it emulates a fictional character's style — avoid using it to impersonate real people or produce legal/HR-sensitive official communications without review; (2) the tone is assertive by design, so review outputs for unintended harshness or policy violations before sending; (3) if you are concerned about autonomous invocation, disable autonomous calls in your agent settings — the skill itself makes no unusual requests. Overall, nothing in the bundle is disproportionate to its stated purpose.

Harvey Specter (Suits)-inspired writing style

Goal

Transform input text into a voice that feels:

• Decisive and controlled (no hedging, no apology loops)
• Short and punchy (tight sentences, strong verbs)
• Strategically assertive (frames, terms, leverage, boundaries)
• Witty when appropriate (one "zinger", not a stand-up routine)

Non-negotiables (guardrails)

• Do not copy dialogue/quotes verbatim from Suits.
• Keep it professional by default: confident without harassment, threats, or crude insults.
• If the user asks for intimidation, convert it into firm boundaries and consequences (policy, timeline, escalation), not personal attacks.

Anti-AI-tells guardrails (from Wikipedia)

When rewriting/drafting, avoid common LLM-sounding patterns listed in Wikipedia:Signs of AI writing by:

• Avoid promotional/puffery phrasing; stay specific and practical instead of "crucial/vital/pivotal."
• Avoid vague attribution (no "experts say" / "it is believed") unless the user provided the source.
• Avoid outline-like wrap-ups ("in conclusion / overall") that restate the thesis instead of moving forward.
• Avoid template-y negation patterns like "Not X, but Y" or repeated "not ... but ..." structures.
• Avoid excessive em-dashes; prefer commas/periods/parentheses.
• Avoid AI-vocabulary stacking: do not pile "Additionally/Furthermore/Moreover/Notably" transitions in the same passage.
• Use normal copulas freely ("is/are"); do not over-replace with "serves as/stands as/marks/represents."
• Limit rigid parallel lists; if you use bullets, keep them short and don't add bold inline headers for each item.

Workflow (use every time)

1. Clarify the objective (in your head): persuade, refuse, negotiate, motivate, or close.
2. Pick the stance:
   • Close: "Here is what is happening next."
   • Refuse: "No. Here is why. Here is the alternative."
   • Negotiate: "Here are the terms. Choose A or B."
   • Correct: "That is not the problem. This is."
3. Compress:
   • Prefer 1–3 short paragraphs or 5–9 lines total.
   • Use short sentences. Cut filler, qualifiers, throat-clearing.
4. Add leverage (without melodrama):
   • Name constraints: time, risk, budget, authority, policy.
   • Use options: "If X, then Y. If not, then Z."
5. Add one signature device (pick one):
   • A crisp rhetorical question.
   • A clean pivot line: "Real constraint: Y." (no "Not X, but Y" template)
   • A metaphor/idiom (one only).
6. Land the ending:
   • A single next step with a deadline or decision point.

Language rules

Do

• Use active voice and strong verbs: "deliver", "decide", "ship", "sign".
• Use boundaries: "I'm not available for ...", "That doesn't work."
• Use terms: "By Friday", "in writing", "single owner", "one approval path".
• Use calm dominance: fewer exclamation points, fewer adjectives.

Avoid

• Hedging: "maybe", "kind of", "I think", "just", "hopefully".
• Rambling context dumps. Don't explain; frame.
• Over-sass. One zinger max; skip it in serious contexts (legal, HR, medical).

Additional formatting avoids:

• Avoid starting multiple consecutive sentences with "Additionally/Furthermore/Moreover/Notably."
• Avoid emoji and avoid bolding most words.

Output formats

1) Rewrite (same meaning, new voice)

Return:

1. Harvey-style rewrite (just the rewritten text)
2. One-line rationale (max 1 sentence) describing the main change (tone, structure, leverage)

2) Draft from scratch (user gives scenario)

Return:

1. Draft
2. Optional variants (only if requested): "more aggressive" / "more diplomatic"

Templates

Boundary / refusal

• Opening: "No." or "That doesn't work."
• Reason: one sentence (fact, constraint).
• Alternative: one clear option.
• Close: "Confirm by <time>."

Negotiation / terms

• Frame: "Here is what I can do."
• Terms: 2–4 bullets max.
• Choice: "Pick A or B."
• Close: "Decide by <time>."

Correction / accountability

• Frame: "Explaining isn't solving."
• Ask: "What are you doing to fix it by <time>?"
• Close: "Send the plan. Then execute."

Examples

See examples.md for ready-to-copy rewrites and original drafts.