Security audit
Security checks for vulnerabilities and agentic risk
This is a Markdown-only OpenAlex guidance skill that coherently teaches scholarly metadata querying and does not include executable code, hidden persistence, credential collection, or destructive behavior.
Install this only if you want your agent to use OpenAlex for scholarly metadata. Review the separate OpenAlex MCP server before enabling runtime tool calls, use a non-sensitive email for OPENALEX_MAILTO, and avoid treating OpenAlex as a source for non-academic web facts or paywalled full text.
The 'When' section lists example triggers such as 'how has research on X grown' and 'publications per year' as activation conditions, but it does not define whether these are exhaustive, what context they apply in, or any negative examples. Because this is a markdown file, such broad phrasing could overlap with ordinary requests and cause unintended invocation.
The README states that the skill 'does not run anything,' implying a passive documentation-only role. However, later text says loading the skill causes the agent to call specific MCP tools correctly, so the documentation minimizes the skill's operational effect on agent behavior.
The activation examples "who is X / what has X published / how cited is X / what's X's most influential work" overlap with common natural-language questions and do not specify boundaries such as academic researchers only, required identifiers, or exclusion cases. Without narrower trigger scope or negative examples, the skill could be invoked for general biographical or non-scholarly queries unintentionally.
No suspicious patterns detected.
