Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only go-to-market writing skill with no code execution or persistence, but users should verify any assumed marketing claims before publishing them.
Install only if you want a drafting aid for GTM materials. Treat inferred differentiators, competitors, metrics, and proof points as draft assumptions and confirm them before using the output in sales, legal, investor, or public marketing materials.
The skill explicitly tells the model to infer specific product details and deliver a full GTM pack even when source facts are missing. In a marketing context, this can produce fabricated differentiators, competitors, or claims that users may treat as real, creating business, legal, and reputational risk if published externally.
This section instructs the model to generate 'specific and evidence-backed' proof points even when no data was provided, merely marking them as assumed. That is dangerous because it encourages authoritative-sounding but unverified marketing evidence, which can be mistaken for substantiated claims in sales, regulatory, or customer-facing materials.
This line says that when no data is provided, the skill should infer a realistic proof point and mark it '(assumed)' and 'never leave a bare placeholder.' But later quality checks require 'flagged placeholders' for missing proof points, which conflicts with the earlier instruction against placeholders. That creates contradictory operator guidance about intended output behavior.
The skill says it will always produce 'all four sections below in order' and those sections are Positioning Statement, Messaging Pillars, Feature & Functionality List, and Use Cases. However, the anti-patterns say 'Do not skip the "not for" section,' implying the skill should include a section that is not actually part of the described output structure.
No suspicious patterns detected.
