Security audit
Security checks for vulnerabilities and agentic risk
This is a simple local detox tracking skill with some broad trigger phrases, but no hidden code, network behavior, or disproportionate permissions were found.
Before installing, be aware that this skill is for personal wellness tracking and may capture symptom or diet information. Use it only for data you are comfortable keeping locally, and consider more specific trigger wording if accidental activation would bother you.
The phrase "start detox" can shadow or conflict with the built-in command pattern around "start," creating ambiguous command resolution. If the platform prioritizes the skill unexpectedly, users may trigger this skill when they intended a native action, reducing reliability and potentially surfacing unwanted health-related workflows.
The trigger phrase "start detox" is broad enough to match ordinary user requests about beginning a detox rather than an explicit request to invoke this specific skill. That can cause accidental activation and route users into unrequested health-related guidance or tracking flows, which is more sensitive because the topic concerns wellness and potentially restrictive diets.
The trigger "detox day" is ambiguous and underspecified, making it likely to collide with natural language in routine wellness discussions. This increases the chance of unintended skill activation, especially in a health context where users may mention detox progress conversationally without wanting the skill to take over.
I do not see a clear natural-language policy violation under the provided rule, because the file does not force a specific language or locale. Therefore this should not be treated as a finding.
No suspicious patterns detected.
