Security audit
Security checks for vulnerabilities and agentic risk
This skill is only a short Markdown instruction, but it frames itself as an autonomous builder across code, infrastructure, and organizations without clear limits or user approval points.
Review this skill carefully before installing. It does not contain executable malware, but its instructions are broad enough that an agent could treat it as permission to help with impactful code, infrastructure, or organizational changes unless you impose your own approval and scope limits.
The skill metadata uses extremely broad, unconstrained invocation language such as 'build' and describes creating code, infrastructure, and organizations without any stated scope limits, approval boundaries, or trigger conditions. This increases the risk that an agent may invoke or route to the skill for high-impact tasks too generally, leading to unsafe autonomous actions or overbroad authority.
The body of the skill advertises autonomous synthesis across software, infrastructure, and organizational domains but provides no warning that these actions can cause costly, irreversible, or security-sensitive changes. In this context, missing warnings are dangerous because the skill explicitly spans multiple high-impact operational areas where users and orchestration systems need clear notice, human review requirements, and safety boundaries.
No suspicious patterns detected.
