The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl and standard shell scripting to perform network requests and parse responses. This is the intended primary purpose of the skill for performing security reconnaissance on Zimbra targets.
[SAFE]: While the skill performs offensive security actions (user enumeration, CVE scanning, SSRF probing), these behaviors are explicitly described in the documentation and match the user's intent when using a 'recon' and 'attack' themed skill. The script does not exfiltrate sensitive local data, does not include obfuscated code, and does not maintain persistence on the agent's system.

zimbra-attack