knoxhub-blog — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited knoxhub-blog (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.
Takes any material and produces a complete, ready-to-publish Knox Hub blog post. Human-written tone. SEO+GEO optimized. All editor fields filled.
The user sends raw material. You deliver a finished post.
The user will send one or more of:
If a URL is provided, fetch it with WebFetch to get the full content. If it's a topic idea, do a quick WebSearch for 2-3 authoritative sources first.
Write a brief analysis (2-3 sentences):
Input type: [URL / text / topic / competitor]
Core topic: [what this is about]
Target audience: [who reads Knox Hub — founders, builders, automators]
Angle: [unique spin we'll take]These rules make the content undetectable by AI detectors (GPTZero, Originality.ai, etc.) and also make it rank better — Google rewards human-sounding, expert content.
#### Rule 1: Burstiness — Break the rhythm AI writes sentences of similar length. Humans don't.
DO THIS: Mix very short sentences (3-5 words) with long complex ones. Wrong: "This tool is useful. It helps automate tasks. You can save time with it." Right: "This tool saves you hours. Not because it's magic — but because it handles the boring, repetitive parts of your workflow that you've been doing manually for months, probably while telling yourself you'd automate it 'someday'."
Rule: Every 3-4 sentences, drop one that's under 6 words. Hard stop.
#### Rule 2: Perplexity — Unpredictable word choices AI predicts the most likely next word. Beat it by choosing words it would never pick.
DO THIS:
#### Rule 3: E-E-A-T — Personal experience signals AI has no personal experience. Google knows this. Fix it.
DO THIS: Add first-person experience markers throughout:
At least 2-3 experience markers per 500 words.
#### Rule 4: Kill AI marker phrases — zero tolerance Scan every sentence. Delete these immediately:
BANNED phrases (EN):
BANNED phrases (RU equivalent):
#### Rule 5: The Sandwich Method Never output raw AI text. Every post gets this treatment:
#### Rule 6: Format unpredictably
Every Knox Hub post follows this flow:
Output in this exact format, ready to copy-paste:
[H1 title — clear, specific, benefit-driven. 50-65 chars. No clickbait.][2-3 sentences. What the post is about + what the reader gets. 120-160 chars. Used as og:description if no meta description set.][Article | Guide | Template | Claude Skill | n8n Template][5-10 relevant tags from Knox Hub taxonomy: automation, n8n, ai, claude, claude-code, claude-skills, lead-generation, saas, business, content, mcp, vibe-coding, productivity, etc.][Optimized title for Google. Include primary keyword. Different from H1 if needed.]_Character count: X/60_
[Compelling description with primary keyword. Clear benefit. CTA if space allows.]_Character count: X/160_
[3-5 keywords: primary keyword, 2-3 long-tail variations. Comma separated.][Descriptive alt text for the cover image. Include primary keyword naturally.]_(Google displays these as rich snippets. AI search engines use them for answers.)_
Q1:
[Question — phrase as users actually search it]A1:
[Answer — 2-4 sentences. Self-contained. Factual. 50-100 words. Directly answers the question.]Q2:
[Question]A2:
[Answer]Q3:
[Question]A3:
[Answer]_(Add Q4, Q5 if the topic warrants it — max 5 FAQs)_
_(Enable "Show Table of Contents" in editor. Auto-generates from H2s, but list them here for reference.)_
1. [H2 section 1 title]
2. [H2 section 2 title]
3. [H2 section 3 title]
4. [H2 section 4 title]
5. [H2 section 5 title]_(Optional. Use for posts where a quick-reference widget adds value — tool lists, key stats, checklist.)_
<div class="sidebar-widget">
<h3>[Widget Title — e.g., "Quick Reference" or "Tools Used"]</h3>
<ul>
<li><strong>[Label]</strong> [Value]</li>
<li><strong>[Label]</strong> [Value]</li>
<li><strong>[Label]</strong> [Value]</li>
</ul>
</div>_(Or write "SKIP — single column layout preferred for this post")_
[Full article content in Markdown.
Rules:
- Start directly — no "In this article we will..." intro
- H2 for main sections, H3 for subsections
- Bold key terms and important points
- Use bullet lists for steps, tool lists, comparisons
- Use numbered lists for sequential steps
- Include at least 1 code block if the topic involves code/templates/workflows
- Add [internal link suggestions] in brackets where relevant Knox Hub posts exist
- No H1 in the content — the Post Title field handles H1
- End with a "Next Steps" or "What to Do Now" section
- Last paragraph: point to Knox Community or related content naturally
]Before presenting output, verify:
---
Ready to publish. Copy each section into the Knox Hub editor.
Suggested cover image prompt for generation:
[1-2 sentence prompt for creating a relevant cover image — describe the visual concept]
Want me to:
- Adjust the tone (more expert / more casual)
- Expand any section
- Write a shorter version
- Generate social captions for this post (/social-captions)_(Keep posts aligned with these topics)_
_(Link to these posts when the topic connects)_
/hub/post/how-to-stop-re-explaining-yourself-to-claude-every-day/hub/post/magic-mcp-turns-claude-code-into-a-ui-designer/hub/post/stop-building-n8n-workflows-by-hand/hub/post/context-engineering-why-your-ai-builds-break/hub/post/master-claude-skills-complete-guide/hub/post/building-automation-stack-from-zero~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.