Glin Profanity MCP

MCP server providing profanity detection tools for AI assistants. Use when reviewing batches of user content, auditing comments for moderation reports, analyzing text for profanity before publishing, or when AI needs content moderation capabilities during workflows.

Security Scan

The skill's description matches a profanity-detection tool, but the runtime instructions rely on npx to fetch and execute code from npm (not declared in requirements), and the skill advertises user-tracking tools without explaining storage, credentials, or privacy — these mismatches warrant caution.

Purpose & Capability

Name/description and the listed tools (profanity detection, censoring, batch operations, context analysis) are coherent with a moderation MCP server. However, the SKILL.md's installation instructions require running 'npx' to fetch the npm package, yet the skill metadata lists no required binaries — that's an omission. Also, features under 'User Tracking' (track_user_message, get_user_profile, get_high_risk_users) imply persistent storage or identifiers; the skill declares no config paths, storage backends, or credentials, which is inconsistent with those capabilities.

Instruction Scope

SKILL.md instructs the agent to add an MCP entry that runs 'npx -y glin-profanity-mcp'. That will cause the agent environment to fetch and execute remote npm package code at runtime. The instructions do not specify where tracked user data is stored, what identifiers are required, retention policies, or where telemetry/requests are sent — the tracking utilities broaden scope without operational detail. There are no instructions that read arbitrary local files, but the guidance is vague about data flows and persistence.

Install Mechanism

There is no formal install spec in the metadata, but the provided installation steps rely on npx, which downloads and runs the package from npm at runtime. Using 'npx -y' implicitly executes remotely fetched code and accepts defaults; that elevates risk compared to a packaged, audited install. The SKILL.md cites npm and GitHub URLs for the project, which helps traceability, but npx-based execution still allows remote code to run and to change if the npm package is updated.

Credentials

The skill declares no required environment variables or credentials, which would be reasonable for a purely local tool — but the presence of user-tracking tools implies some form of storage, identifiers, or external service (DB, analytics, or remote API). The absence of declared config paths, storage endpoints, or credentials (for example, for a database) is a mismatch. There is also no privacy or data-handling guidance in the SKILL.md.

Persistence & Privilege

always is false (good). The skill runs via npx when invoked by the agent; it does not request explicit persistent platform privileges in the metadata. However, because the runtime uses npx to fetch and execute remote code and the skill provides user-tracking features, the effective blast radius is higher if the agent invokes it autonomously (default behavior). Consider limiting autonomous invocation until you audit the package.

What to consider before installing

This skill looks like a reasonable profanity-moderation tool, but exercise caution before installing. The SKILL.md instructs the agent to run 'npx -y glin-profanity-mcp', which will download and execute code from npm/GitHub at runtime — verify the npm package and repository (look at the exact package contents and history). Ask the author where user tracking data is stored, what identifiers are collected, retention/ deletion policies, and whether any external endpoints are contacted. Prefer a pinned package version (not 'npx -y' which fetches the latest), run it in a sandbox first, and avoid enabling autonomous invocation until you've reviewed the source and data-handling details.

Glin Profanity MCP Server

MCP (Model Context Protocol) server that provides profanity detection as tools for AI assistants like Claude Desktop, Cursor, and Windsurf.

Best for: AI-assisted content review workflows, batch moderation, audit reports, and content validation before publishing.

Installation

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "glin-profanity": {
      "command": "npx",
      "args": ["-y", "glin-profanity-mcp"]
    }
  }
}

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "glin-profanity": {
      "command": "npx",
      "args": ["-y", "glin-profanity-mcp"]
    }
  }
}

Available Tools

Core Detection

Tool	Description
check_profanity	Check text for profanity with detailed results
censor_text	Censor profanity with configurable replacement
batch_check	Check multiple texts at once (up to 100)
validate_content	Get safety score (0-100) with action recommendation

Analysis

Tool	Description
analyze_context	Context-aware analysis (medical, gaming, etc.)
detect_obfuscation	Detect leetspeak and Unicode tricks
explain_match	Explain why text was flagged
compare_strictness	Compare detection across strictness levels

Utilities

Tool	Description
suggest_alternatives	Suggest clean replacements
analyze_corpus	Analyze up to 500 texts for stats
create_regex_pattern	Generate regex for custom detection
get_supported_languages	List all 24 supported languages

User Tracking

Tool	Description
track_user_message	Track messages for repeat offenders
get_user_profile	Get user's moderation history
get_high_risk_users	List users with high violation rates

Example Prompts

Content Review

"Check these 50 user comments and tell me which ones need moderation"
"Validate this blog post before publishing - use high strictness"
"Analyze this medical article with medical domain context"

Batch Operations

"Batch check all messages in this array and return only flagged ones"
"Generate a moderation audit report for these comments"

Understanding Flags

"Explain why 'f4ck' was detected as profanity"
"Compare strictness levels for this gaming chat message"

Content Cleanup

"Suggest professional alternatives for this flagged text"
"Censor the profanity but preserve first letters"

When to Use

Use MCP server when:

• AI assists with content review workflows
• Batch checking user submissions
• Generating moderation reports
• Content validation before publishing
• Human-in-the-loop moderation

Use core library instead when:

• Automated real-time filtering (hooks/middleware)
• Every message needs checking without AI involvement
• Performance-critical applications (< 1ms response)

Resources

• npm: https://www.npmjs.com/package/glin-profanity-mcp
• GitHub: https://github.com/GLINCKER/glin-profanity/tree/release/packages/mcp
• Core library: https://www.npmjs.com/package/glin-profanity