dbs-ai-check

dbs-ai-check is an AI writing fingerprint detection skill designed to analyze text for patterns commonly associated with AI-generated writing. Instead of simply labeling content as "AI-written," the skill produces a structured diagnostic report that identifies specific writing behaviors and stylistic signals in context. Its philosophy is that so-called "AI flavor" comes from writing that feels overly polished, uniform, and frictionless rather than from poor quality. By highlighting these signals directly in the original text, the tool helps users better understand how their writing reads to others without automatically rewriting or masking those patterns.

The skill scans submitted content against 22 predefined AI writing characteristics, such as excessive rhetorical symmetry, overuse of contrast structures like "not X but Y," artificial emotional smoothness, repetitive concession templates, and over-optimized philosophical framing. Reports are generated in reading order rather than grouped by category, with each flagged passage quoted directly and accompanied by a concise explanation and severity level. By default, the skill only diagnoses issues and avoids rewriting. If the user explicitly requests revision guidance, the skill enters a guided rewrite mode that asks targeted questions about the writer's actual intent before suggesting alternative directions.

This skill is useful for writers, editors, creators, marketers, students, and content teams who want to evaluate whether their writing feels overly synthetic or formulaic. It is particularly suited for long-form essays, social media posts, newsletters, scripts, and AI-assisted drafts where authenticity and personal voice matter. The skill is also designed to reduce false positives by recognizing that many rhetorical devices commonly associated with AI writing are also widely used by human authors depending on genre and context.