compliance-check_skill

Overview

This skill performs pre-post compliance checks against multiple free, continuously updated wordlists so content is only published when safe. It is ready to use out of the box and does not require users to build or maintain their own dictionaries. The focus is fast, reliable detection of sensitive, prohibited, or advertising-claim issues before publishing.

How this skill works

It compares user-provided text to a suite of built-in and synced free wordlists covering profanity, prohibited terms, advertising violations, pornographic content, political/violent flags, and more. The skill returns structured JSON (pass, short report, summary, hits by source, compliance tips) and optionally a human-readable Markdown report. Users can run checks on raw strings or files; no separate synchronization or manual dictionary addition is required.

When to use it

• Before posting user-generated content to ensure compliance
• When reviewing drafts for sensitive or prohibited terms
• To screen marketing copy for advertising law violations or exaggerated claims
• As a quick content moderation step prior to listing or publishing
• When you need a fast, plug-and-play sensitive-word scan without building lists

Best practices

• Run checks on both the final draft and key variations (titles, descriptions, captions)
• Treat results as guidance: review hits manually to decide context and intent
• Use the JSON output for automation and the report format for human review
• Optionally add a small custom list for platform-specific rules via config/sensitive_words.txt
• Log checks and hits for auditability and to refine content workflows

Example use cases

• Scan a product listing title and description before publishing to marketplaces
• Check social media posts or ad copy for exaggerated or medical claims
• Screen user-submitted articles, comments, or forum posts for profanity or sensitive topics
• Automate pre-release checks in a CI pipeline that blocks non-compliant drafts
• Provide moderators with a quick report that highlights problematic phrases and suggested fixes