secure-skill-content-sanitization

Detect and neutralize hidden content in markdown, HTML, SVG, and text files.

3|Updated Apr 5, 2026

Search Tags:#markdown #html #security #sanitization #unicode #prompt-injection #content-scanning

Ready to run

npx skills add https://github.com/dvy1987/agent-loom --skill secure-skill-content-sanitization

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: secure-skill-content-sanitization
Source: https://github.com/dvy1987/agent-loom/tree/main/.agents/skills/secure-skill-content-sanitization
Command: npx skills add https://github.com/dvy1987/agent-loom --skill secure-skill-content-sanitization

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prevents invisible or obfuscated content in markdown, HTML, and text files from silently influencing AI agents by detecting and neutralizing CSS-hidden text, HTML comments and interactive blocks, zero-width and bidi characters, homoglyphs, and misleading links before content enters agent context.

Core Features & Use Cases

Preprocessing Sanitization: Strips HTML, extracts and scans comments, expands collapsed content, and normalizes Unicode to eliminate hidden instruction channels.
Attack Detection: Flags CSS-hidden elements, javascript/data links, image-based exfiltration, zero-width characters, bidi overrides, and homoglyph obfuscation with severity levels (CRITICAL / HIGH).
Use Case: Run during a repository skill audit or CI pre-commit to block or require review for any file containing invisible instructions or link-based exfiltration attempts.

Quick Start

Run a repository sanitization: sanitize the README.md and all .md/.html files, remove zero-width characters, extract comments for review, and produce a findings report.

Frequently Asked Questions about secure-skill-content-sanitization

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I sanitize markdown and HTML files to prevent prompt injection attacks?▼

Sanitizing markdown and HTML against prompt injection involves detecting and neutralizing visually hidden content like CSS-hidden text, HTML comments, and malicious links before it enters an agent's context. This process strips HTML, extracts comments, and normalizes Unicode to eliminate hidden instruction channels.

What is hidden text in markdown and how does it affect AI agents?▼

Hidden text in markdown uses CSS-hidden elements, HTML comments, or zero-width and bidi characters to silently influence AI agents. This obfuscated content acts as a hidden instruction channel, bypassing human review to inject malicious or misleading directives directly into agent context.

How do I detect zero-width characters and homoglyphs in repository files?▼

Detecting zero-width characters and homoglyphs in repository files requires normalizing Unicode and scanning for obfuscation techniques across markdown, HTML, SVG, and text files. The scan flags these invisible characters with severity levels to prevent misleading or malicious content from entering agent context.

Can I scan for image-based exfiltration and javascript links in SVG files?▼

Yes, you can scan SVG and HTML files to flag image-based exfiltration and javascript or data links. The sanitization process identifies these attack vectors alongside CSS-hidden elements and bidi overrides, categorizing them with CRITICAL or HIGH severity levels in a standardized findings report.

Does content sanitization work for CI pre-commit repository audits?▼

Content sanitization works for CI pre-commit repository audits by scanning all markdown and HTML files to block or require review for any file containing invisible instructions or link-based exfiltration attempts. It produces a standardized sanitization report to enforce repository security.

What is the best way to extract and review HTML comments in markdown before processing?▼

The best way to extract and review HTML comments in markdown is to use a preprocessing sanitization step that strips HTML, expands collapsed content like <details> blocks, and extracts comments for manual review before the content enters an agent's context.