Security audit
Security checks for vulnerabilities and agentic risk
This markdown-only side-hustle planning skill has some language and metadata quality issues, but it does not include code, persistence, credential access, or hidden high-impact behavior.
Before installing, expect a simple advice-generation skill for Korean side-hustle ideas. Review the mixed-language wording and broad trigger behavior if precise Korean-only output matters, and independently verify legal, tax, platform, and income claims before acting on the advice.
The activation description is broad enough to trigger on many common requests about ideas, income, or side jobs, which can cause the skill to activate when the user did not clearly ask for this specific capability. Over-broad activation increases the chance of unintended routing, irrelevant responses, and policy bypass of more appropriate skills or base-model behavior.
The skill metadata mixes Korean with Chinese and Russian text without giving the user a language choice, which can push multilingual output unexpectedly. This can confuse users, degrade reliability, and create prompt-routing or instruction-following issues if downstream components assume a single language.
The operational instructions continue the mixed-language pattern, increasing the likelihood that generated responses will contain unexpected Chinese or other non-requested language fragments. In context, this makes the skill less trustworthy and can mislead users about requirements, steps, or financial estimates in a side-income planning workflow.
The principles section reinforces multilingual wording without justification, which perpetuates the same output-integrity problem throughout the skill. While less severe than broad activation or core instructions, it still contributes to inconsistent behavior and user confusion.
