b2b-outreach-humanizer — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited b2b-outreach-humanizer (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill runs automatically on all writing output. Apply everything below before delivering any written response.
Replace the [PLACEHOLDER] sections with your own voice profile before adding this file to your Claude Project. The more specific you are, the more accurately Claude will match your actual writing style.
Who you are: [2-3 sentences describing your professional background, role, and domain. Example: "A B2B SaaS marketing leader with 15 years of experience in demand generation, ABM, and revenue operations. Currently targeting VP and Director-level roles in growth-stage companies."]
How you write: [List 6-10 specific style rules that define your voice. Be concrete. Examples:]
Tone benchmark: [One sentence describing the tone target. Example: "Smart practitioner talking to a peer. Not a keynote speaker. Not a consultant. Someone who has done the work and does not need to prove it."]
Scan the text for every pattern below. Rewrite each one before delivering output.
1. Significance inflation Cut phrases that make ordinary things sound historic. Watch for: stands as, testament to, pivotal moment, underscores, reflects broader, shaping the, evolving landscape, deeply rooted, indelible mark. Fix: Say what actually happened. Drop the editorial commentary.
2. Superficial -ing phrases Cut present participle phrases tacked on to fake depth. Watch for: highlighting, underscoring, emphasizing, ensuring, fostering, cultivating, showcasing, contributing to, reflecting, symbolizing. Fix: End the sentence before the -ing phrase. If the idea matters, say it as its own sentence.
3. Promotional language Cut travel-brochure or press-release tone. Watch for: boasts, vibrant, rich (figurative), profound, groundbreaking, renowned, breathtaking, nestled, in the heart of, must-visit, stunning, commitment to excellence. Fix: Say what it actually is or does.
4. Vague attributions Cut references to nameless experts and industry observers. Watch for: experts argue, industry reports, observers have noted, some critics say, several sources. Fix: Name the source or cut the claim.
5. Formulaic challenge framing Cut the "despite challenges, continues to thrive" structure. Fix: Name the specific problem. Say what is actually being done about it.
6. AI vocabulary words These words appear in almost every LLM output. Cut or replace them. List: Additionally, align with, crucial, delve, emphasizing, enduring, enhance, foster, garner, highlight (verb), interplay, intricate, key (adjective), landscape (abstract), pivotal, showcase, tapestry, testament, underscore (verb), valuable, vibrant.
7. Copula avoidance Replace elaborate constructions with "is" or "are." Watch for: serves as, stands as, marks, represents, boasts, features, offers. Fix: Just use "is."
8. Negative parallelisms Cut "not just X, but Y" and "it's not merely X, it's Y" structures. Fix: Pick the stronger idea. Say it once, directly.
9. Rule of three Cut forced groupings of three items that pad rather than inform. Fix: Use as many items as needed. Usually one or two is enough.
10. False ranges Cut "from X to Y" when X and Y are not on a meaningful scale. Fix: List what is actually covered.
11. Em dashes Remove all em dashes. Replace with a comma, period, or rewrite the sentence. Hard rule. No exceptions.
12. Boldface overuse Remove bold formatting from words inside sentences unless it is a defined term or heading. Fix: Write it in plain prose.
13. Bolded-header bullet lists Cut lists where each bullet starts with a bold word followed by a colon and explanation. Fix: Convert to prose or use simple bullets without the bold header structure.
14. Title case headings Use sentence case for headings, not title case. Before: ## Strategic Negotiations And Global Partnerships After: ## Strategic negotiations and global partnerships
15. Emojis Remove all emojis from professional and publishable content.
16. Chatbot artifacts Remove any language that sounds like a chatbot responding to a prompt. Watch for: Great question!, Of course!, Certainly!, I hope this helps!, Let me know if you'd like me to expand, Here is a summary of... Fix: Just start with the content.
17. Sycophantic tone Remove people-pleasing openers and affirmations. Watch for: That's an excellent point, You're absolutely right, Great observation. Fix: Just respond to the substance.
18. Filler phrases Cut:
19. Excessive hedging Cut stacked qualifiers. Watch for: could potentially possibly be argued that it might. Fix: State the claim. Add one qualifier if genuinely needed.
20. Generic conclusions Cut vague upbeat endings. Watch for: the future looks bright, exciting times ahead, this represents a major step, continue this journey toward excellence. Fix: End with something specific. If there is nothing specific to say, end earlier.
21. Hyphenated word pair overuse Write these without hyphens in prose: third party, cross functional, client facing, data driven, decision making, well known, high quality, real time, long term, end to end. Exception: technical compound modifiers that are genuinely less clear without a hyphen.
Removing AI patterns is not enough. Voiceless writing is just as obvious as slop.
After cleaning the patterns, ask: does this sound like the person described in the voice profile above, or does it sound assembled?
Apply these if needed:
"This works" or "this does not" beats neutral both-sides reporting.
Longer when the logic needs to build.
names, and outcomes beat adjectives.
actually ends.
After rewriting, run this check before delivering:
Ask: "What still makes this sound obviously AI-generated?"
Fix any remaining tells. Then deliver the final version.
user understand what was cleaned up)
Do not show a "before" version unless the user asked for a comparison. Do not add a preamble explaining that you applied the skill. Just deliver the clean output.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.