Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward document-analysis helper, with some overly broad activation phrases users should be aware of.
Install this only if you intend to use Akashic MCP services for document processing, since document contents may be handled by those tools. Be careful with sensitive files, and prefer invoking it only for explicit document-analysis tasks until the publisher narrows the broad triggers.
The trigger "read pdf" overlaps with the built-in verb "read," creating a shadowing/conflict risk where this skill may intercept requests intended for a more general or trusted built-in capability. In a skill that processes documents via external tools, such interception can alter execution flow and send content into a different handling path than the user expected.
The trigger phrase "extract from" is overly broad and can match many unrelated user requests, causing the skill to activate outside its intended document-analysis context. Because this skill can process user-supplied files and invoke external MCP tools, accidental invocation increases the chance of unintended file handling or data exposure through overbroad routing.
The trigger "process file" is highly generic and likely to overlap with many unrelated workflows that involve files, leading to unintended skill selection. In this context, accidental activation could route arbitrary file-processing requests into this skill's document pipeline and associated tools, increasing the risk of misprocessing sensitive content.
No suspicious patterns detected.
