Security audit
Security checks for vulnerabilities and agentic risk
This skill only changes how the agent asks about and controls answer length, with no code execution, data access, network use, or hidden persistence.
Before installing, understand that this skill may ask you to choose answer depth when you use common words like brief or detailed; it does not appear to access private data or run code.
The manifest description lists triggers such as "short version," "brief answer," and "detailed answer," which are common conversational phrases that could appear in many contexts unrelated to this skill. Because these phrases are broad and the manifest does not tightly constrain invocation context there, the skill risks unintended activation.
The guidance says to use the skill when a user mentions "tokens, budget, depth, or response length" or says terms like "brief" and "exhaustive," but these conditions are broad and could match ordinary requests. Although there is one exclusion at L17, the file still lacks sufficiently specific constraints or negative examples to reliably separate intentional activation from normal conversation.
