Stats
Actions
Tags
From journalism-toolkit
Detects AI-generated text, click-farm content, and synthetic articles via multi-tool consensus (GPTZero, Originality.ai, Copyleaks), perplexity/burstiness analysis, stylometric fingerprinting, and pattern recognition.
How this agent operates — its isolation, permissions, and tool access model
Agent reference
journalism-toolkit:agents/journalism/ai-content-detectorThe summary Claude sees when deciding whether to delegate to this agent
**Purpose**: Detect AI-generated text, click-farm content, synthetic articles **Outputs**: AI probability scores, detection consensus reports, stylometric fingerprints **When to Use**: Suspected AI-generated or click-farm text content (Tier 2 - conditional) **Key Tools**: GPTZero, Originality.ai, Copyleaks, perplexity/burstiness analysis, stylometric fingerprinting - Multi-tool AI text detectio...
Purpose: Detect AI-generated text, click-farm content, synthetic articles Outputs: AI probability scores, detection consensus reports, stylometric fingerprints When to Use: Suspected AI-generated or click-farm text content (Tier 2 - conditional) Key Tools: GPTZero, Originality.ai, Copyleaks, perplexity/burstiness analysis, stylometric fingerprinting
npx claudepluginhub swarochish/journalism-toolkitUse this agent when you need to analyze text for signs of AI or LLM-generated content. This includes reviewing articles, essays, documentation, or any written content to identify telltale patterns of machine-generated text. The agent evaluates linguistic patterns, structural elements, and stylistic markers that commonly indicate AI authorship. <example> Context: User wants to verify if a submitted article was human-written. user: "Can you check if this blog post about climate change was written by AI?" assistant: "I'll use the ai-writing-detector agent to analyze this text for signs of AI generation." <commentary> Since the user is asking to analyze text for AI authorship, use the Task tool to launch the ai-writing-detector agent to perform a comprehensive analysis. </commentary> </example> <example> Context: User is reviewing content before publication. user: "Here's a draft article from a contributor. I want to make sure it's genuinely human-written before we publish." assistant: "I'll launch the ai-writing-detector agent to examine this draft for AI writing indicators." <commentary> The user needs content verification before publication, making this an ideal use case for the ai-writing-detector agent. </commentary> </example> <example> Context: User is editing their own writing and wants to avoid sounding like AI. user: "I wrote this myself but I'm worried it sounds too 'AI-like'. Can you analyze it?" assistant: "Let me use the ai-writing-detector agent to identify any passages that might read as AI-generated, so you can revise them." <commentary> Even for human-written content, the ai-writing-detector agent can identify patterns that unintentionally mirror AI writing styles. </commentary> </example>
Detects AI-sounding patterns and suspect zones in KO/EN/ZH/JA text using pattern packs, stylometric burstiness/MATTR analysis, and AI-lexicon density. Read-only and never rewrites.
Detects AI-written signatures in Korean text by scanning for 10 categories × 40+ sub-patterns, outputting span-level JSON findings with severity, offset, reason, and suggested fix.