Security audit
Security checks for vulnerabilities and agentic risk
This is a local plain-text reference skill with no credential, network, persistence, or mutation behavior, though some finance content is generic or mismatched.
Install only if you want a lightweight local reference stub. Do not rely on it for trading, regulatory, or risk decisions without independent authoritative sources, because the included finance material is generic and partly off-topic.
The skill description is broad and generic enough that an agent may invoke it in situations where the user did not explicitly request this skill. Overly loose activation language increases the chance of unintended tool selection, which can cause prompt-scope confusion or irrelevant skill execution even though this specific skill appears read-only and non-networked.
The command triggers are common single words like 'intro', 'risks', and 'checklist', which can easily appear in ordinary conversation and unintentionally activate the skill. Generic triggers create a realistic risk of false-positive invocation and routing errors, especially in multi-skill environments where common words overlap heavily.
The file header claims this is a quant-finance reference tool, but the embedded text repeatedly uses vague, domain-inconsistent boilerplate. In an agent skill, this kind of deceptive or low-integrity labeling is dangerous because agents may route sensitive finance-related queries to content that is not authoritative for that domain.
The commands in this region present generic operational inspection, equipment, and shutdown guidance that does not match a quant-finance reference tool. This mismatch can mislead downstream agents or users into trusting irrelevant or unsafe instructions, creating prompt-surface integrity issues and increasing the chance of misuse in finance workflows.
No suspicious patterns detected.
