The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): The skill does not access sensitive local files or hardcode credentials. It uses web_search to gather public data from trademark databases through search engine indexing. It explicitly avoids direct automated requests to these sites to prevent being blocked.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): There are no package installations (pip/npm) or remote script downloads. The skill logic is entirely instructional.
[Indirect Prompt Injection] (LOW): The skill processes untrusted data from the web (search snippets). However, it lacks dangerous capabilities like file writing or command execution, limiting the impact of any potential injection. It uses structured output templates to maintain a consistent format.
[Privilege Escalation] (SAFE): No administrative or elevated privilege commands are present.
[Obfuscation] (SAFE): The instructions are clear and do not contain encoded or hidden text.

trademark-search