Security audit
Security checks for vulnerabilities and agentic risk
This skill is a plain Markdown guide for structured medical literature analysis, with no executable code or hidden system-level behavior found.
Install only if you want an agent to help critique and organize scientific papers. Do not provide confidential unpublished research, patient-identifying clinical information, or restricted PDFs unless you are comfortable processing them in your agent environment, and verify any generated citation or related-study suggestions before relying on them.
The skill's trigger language is very broad, covering many common paper-reading and research-assistance requests, which can cause over-activation outside a narrowly intended scope. That increases the chance the agent will route unrelated or sensitive requests into this skill, producing inappropriate handling, unexpected retrieval behavior, or bypassing a better-scoped skill.
The skill accepts PDFs, identifiers, screenshots, and free-form inputs that may contain unpublished, copyrighted, or sensitive research/clinical material, but it provides no explicit notice about retrieval, external lookup, data handling, or privacy implications. Users may unknowingly submit sensitive content or trigger network-based retrieval without informed consent, creating confidentiality, compliance, and trust risks.
The phrase "Offer plugins when genuinely useful" does not define concrete trigger conditions or clear boundaries for when plugin suggestions should or should not appear. This ambiguity can lead to inconsistent or unintended invocation behavior, especially since the file also says to load the file when offering plugin suggestions.
No suspicious patterns detected.
