Security audit
Security checks for vulnerabilities and agentic risk
This is a simple product-planning skill with no executable code or privileged behavior, though its activation wording is broad enough that users may want to invoke it explicitly.
Installers should be aware that this skill may trigger on broad product-help wording; prefer explicit invocation for product validation tasks and consider narrowing triggers, but the reviewed artifact shows no malicious or privileged behavior.
The trigger list includes extremely generic terms such as "ask," "product," "very," "high," and "ready," which are common in ordinary conversations and can cause the skill to activate unintentionally. Over-broad activation increases the chance of prompt routing errors, where this skill intercepts unrelated user requests and influences outputs outside its intended scope.
The description says to use the skill when a user asks for broad categories like creative content, practical workflows, artifacts, analysis, or implementation support, without clearly limiting the product-validation context. This ambiguity can make the skill eligible for many unrelated tasks, increasing misrouting risk and reducing predictability of skill selection.
The default invocation phrase is highly generic and closely matches ordinary user language, which increases the chance that the skill is triggered unintentionally during unrelated conversations. This can cause prompt-routing confusion, invoke the wrong capability without clear user intent, and create opportunities for prompt-injection or unintended data exposure through misrouting.
Enabling implicit invocation without narrow trigger constraints allows the skill to activate based on broad contextual matches instead of clear user consent. In a planning/advice skill with generic product-help semantics, this broadens the attack surface for accidental invocation, prompt hijacking through crafted phrasing, and inappropriate routing of user requests into this skill.
The trigger scope is genuinely overbroad: generic keywords such as "ask," "product," "very," "high," and "quality" can match many unrelated user requests, causing the skill to activate outside its intended domain. This can lead to incorrect routing, unnecessary prompt injection exposure from untrusted content, and degraded reliability because the agent may apply this workflow when a more appropriate skill should handle the request.
No suspicious patterns detected.
