The Agent Skills Directory

[SAFE]: The skill's content is entirely instructional and follows best practices for agent skill development. No prompt injections, obfuscation, or persistence mechanisms were detected.
[DATA_EXPOSURE]: No indicators of sensitive data harvesting or exfiltration were found. The skill interacts with local skill configuration files, which is appropriate for its stated purpose.
[COMMAND_EXECUTION]: The skill uses standard filesystem and shell tools for routine management tasks like listing directories and creating markdown files. There are no attempts to execute arbitrary code or escalate system privileges.
[INDIRECT_PROMPT_INJECTION]: The skill defines workflows for reading and auditing external skill files. This creates a surface where untrusted instructions in audited files could influence the agent; however, the skill's guidelines focus on manual structural auditing and template-based creation, which minimizes the risk of unintended instruction execution.

hermes-skill