Security audit
Security checks for vulnerabilities and agentic risk
This is a copywriting helper skill with no code execution, data access, persistence, or hidden system behavior, though users should be careful with broad triggers and regulated health claims.
Before installing, treat Copyku as a text-generation assistant for marketing drafts. Review generated claims before publishing, especially for health, finance, legal, or other regulated products, and use explicit prompts if you do not want broad keyword detection to shape the output.
The README says users can 'langsung sebutkan kebutuhan' and gives examples in the form 'copyku, ...', but it does not clearly define whether any phrase containing 'copyku' plus a request will activate the skill or what inputs should not trigger it. This ambiguity can cause unintended invocations because the trigger scope and exclusion conditions are not explicitly constrained.
The skill explicitly supports health-industry marketing templates and claims like 'klinis, teruji, dokter rekomendasikan' without prominent warnings about regulated claims, jurisdictional compliance, or review requirements. In a copywriting skill, this increases the risk that users generate promotional text that violates medical advertising rules or makes unsafe, misleading health claims.
The skill says users can activate it by simply stating phrases like "caption produk skincare," "headline kursus online," or "iklan makanan ringan." These are common everyday content requests and the document does not define narrower invocation constraints or exclusion conditions, which could cause unintended activation.
The document states that Copyku automatically detects broad keywords such as "caption," "headline," "story," "jual," and "bulk" and maps them to behaviors. These terms are common in normal conversation, and the skill does not specify when detection should apply versus when it should be ignored.
No suspicious patterns detected.
