Security audit
Security checks for vulnerabilities and agentic risk
This is a plain Markdown photography coaching skill with broad but low-risk activation examples and no evidence of hidden code execution, credential access, persistence, or data exfiltration.
Installers should understand that this skill may activate on broad photography-related phrases, so use explicit wording when you want photography help. It appears suitable as a reference/coaching skill and does not show evidence of hidden system access.
The README suggests very generic trigger phrases like "Build my portfolio" and "Price my photography," which overlap with common user requests and could cause the skill to activate unintentionally in broad conversational contexts. While not directly enabling code execution or data exfiltration, overly broad invocation language can lead to prompt-routing confusion and unintended skill use.
The trigger phrase "Review my portfolio" is broad natural language that could easily appear in ordinary conversation outside an explicit skill-invocation context. In systems that route skills based on phrase matching, this can cause unintended activation and scope confusion, especially when user content or surrounding prompts mention similar wording.
The trigger "Settings for [genre]" is ambiguous because "genre" is underspecified and the phrase structure is common across many domains, not just photography. This increases the chance that unrelated requests are misrouted into the skill or that adversarial prompt content can intentionally collide with the trigger to steer execution.
The trigger "Teach me [technique]" is extremely generic and likely to overlap with many unrelated user requests across other skills or core assistant behavior. Such vagueness makes reliable routing difficult and can lead to unintended activation, priority conflicts, or prompt-injection-style abuse through crafted text that mimics activation phrases.
No suspicious patterns detected.
