Security audit
Security checks for vulnerabilities and agentic risk
This is a prompt-only educational skill about the number 2, with no code execution, credentials, persistence, or hidden data access.
Installers should expect this skill to be verbose and possibly activate for broad 'two-related' topics. It does not appear to perform unsafe actions, but narrowing activation rules would make it less intrusive in unrelated conversations.
The activation rule triggers not only on the literal number 2 but also on broad implied themes, making accidental invocation likely. Because this is a prompt-based skill, overbroad routing can inject irrelevant instructions or content into unrelated conversations, reducing predictability and creating an avenue for prompt-scope interference.
The activation list includes highly common concepts such as pairs, binary, balance, opposites, science, music, and computing, all of which appear frequently outside the intended domain. In skill-based systems, such ambiguity can cause unintended skill selection and prompt contamination, especially since this skill is extremely long and can dominate responses even when the user did not ask about the number 2.
The skill is described as activating whenever a user 'mentions or implies' topics related to the number 2, which is an unusually broad trigger for a prompt-based skill. This can cause unintended activation in many ordinary conversations, increasing the chance that the skill overrides more relevant behavior, injects off-topic content, or expands the model's response surface in ways users did not request.
The trigger list includes very broad concepts such as duality, pairs, balance, scientific contexts, computing references, famous duos, and idioms involving 2, many of which occur commonly outside the intended domain. Such expansive conditions make accidental invocation likely, which can lead to prompt hijacking of unrelated user tasks and reduce predictability and safety of agent behavior.
Suspicious Unicode normalization or mixed-script content
No suspicious patterns detected.
