The Agent Skills Directory

[SAFE]: The skill is a defensive utility designed to sanitize workflows by removing sensitive information like API keys, personal data, and local file paths.
[PROMPT_INJECTION]: The skill ingests untrusted workflow data from users (ingestion point: SKILL.md), which represents an indirect prompt injection surface. No explicit boundary markers are defined. However, the skill has no dangerous capabilities such as subprocess calls or file-writes, and its core purpose is the sanitization and transformation of that data.
[DATA_EXFILTRATION]: No exfiltration patterns were detected. The skill instructions explicitly forbid the inclusion of credentials, tokens, or private data in the output.

hermes-imports