The Agent Skills Directory

[PROMPT_INJECTION]: No patterns of instruction overriding, safety bypass, or role-play injection were detected. The skill includes proactive GDPR compliance instructions when dealing with individual entities.
[DATA_EXFILTRATION]: The skill does not access sensitive local files or credentials. Network activity is limited to standard entity research tools and well-known services such as Google Knowledge Graph and Wikidata.
[OBFUSCATION]: Analysis found no use of Base64 encoding, zero-width characters, or homoglyph-based obfuscation in the instructions or references.
[REMOTE_CODE_EXECUTION]: No mechanisms for downloading or executing remote code or unverifiable dependencies were identified.
[COMMAND_EXECUTION]: The skill does not use dangerous shell commands, privilege escalation (sudo), or persistence mechanisms.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted entity data for auditing. However, it lacks executable capabilities (such as exec, eval, or subprocess) that could be triggered by such data and primarily outputs structured reports.

entity-optimizer