Security audit
Security checks for vulnerabilities and agentic risk
This is a bounded X reply assistant that can like and post from the user's account, and that behavior is disclosed and aligned with its purpose.
Install this only if you are comfortable letting the agent use your logged-in X session to like posts and publish short replies on your behalf. Give a narrow scope and count limit when using it, and ask to review drafts first if you do not want automatic posting.
The instruction 'Use this file whenever drafting a comment, reply, or post' is extremely broad and overlaps with common, everyday writing tasks. It does not define boundaries, exclusions, or specific contexts, so the skill could be invoked in many situations where it was not intended.
The instruction 'Use this file while handling notifications or mentions' is ambiguous because it does not clearly limit the context to X.com until later steps. A user or agent could reasonably interpret it as applying to any notification workflow, increasing the risk of unintended invocation.
The document instructs the agent to always use this file for drafting and defines a required emotional and behavioral baseline, without offering the user a choice in tone or style. This imposes a fixed communication style rather than allowing user preference or opt-in.
No suspicious patterns detected.
