sanitizer

Sanitize external text inputs by removing HTML comments, invisible characters, and GitHub tokens.

Updated Jan 8, 2026
One-click install
npx skills add https://github.com/takemo101/compose-workflow --skill sanitizer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sanitizer
Source: https://github.com/takemo101/compose-workflow/tree/main/.opencode/skill/sanitizer
Command: npx skills add https://github.com/takemo101/compose-workflow --skill sanitizer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill protects your AI from malicious inputs by removing potentially harmful elements like prompt injection attempts, sensitive tokens, and unwanted formatting from external text sources.

Core Features & Use Cases

  • Prompt Injection Defense: Strips HTML comments and invisible characters that could be used to manipulate AI instructions.
  • Token Redaction: Masks sensitive GitHub tokens to prevent accidental exposure.
  • Input Normalization: Cleans up markdown and HTML to ensure safe processing.
  • Use Case: When processing user-submitted GitHub issues or comments, this Skill ensures that any hidden instructions or leaked tokens are neutralized before being fed to the AI, maintaining system security.

Quick Start

Sanitize the content of the file named 'user_input.txt' using the sanitizer script.

Frequently Asked Questions about sanitizer

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

FAQPage Schema
How do I prevent prompt injection when processing GitHub issues with AI?

Preventing prompt injection involves sanitizing external text inputs to strip hidden instructions. This Skill removes HTML comments and invisible characters from user-submitted GitHub issues before feeding them to your AI, neutralizing malicious manipulation attempts.

How do I sanitize external text inputs to remove sensitive GitHub tokens?

To sanitize external text inputs, this Skill masks sensitive GitHub tokens using bash scripts with sed and perl. It identifies and redacts exposed credentials from user-generated content, preventing accidental data exfiltration during AI processing.

What is the best way to clean invisible characters from user-submitted comments?

The best way to clean invisible characters is text manipulation using bash scripts. This Skill removes hidden formatting elements and normalizes markdown and HTML, ensuring that external user-submitted comments are safe for AI processing.

Does this input validation approach work without installing external dependencies?

Yes, this input validation approach works without external dependencies. It relies solely on built-in bash scripts utilizing standard command-line text manipulation tools like sed and perl to sanitize data.

Can I use this sanitization script for processing bulk GitHub comments?

You can use this sanitization script for bulk processing by passing your GitHub comments as file inputs. The script cleans the content by removing HTML attributes, invisible characters, and tokens, preparing safe text outputs for your AI workflows.

What specific HTML attributes are removed to secure AI inputs?

To secure AI inputs, the sanitization process removes specific HTML attributes, HTML comments, and invisible characters. This targeted text manipulation prevents hidden payloads in external content from executing prompt injection or data exfiltration attacks.