The Agent Skills Directory

[SAFE]: The skill uses the Bash tool to interact with a well-known service (Google's Generative Language API at generativelanguage.googleapis.com). This is considered safe under trust-scope rules.
[SAFE]: User-provided prompts and parameters are stored in temporary JSON files before being passed to curl via the -d @file syntax. This is a secure pattern that prevents shell command injection vulnerabilities.
[SAFE]: Sensitive information (GEMINI_API_KEY) is handled through environment variables and the skill includes explicit instructions to avoid logging the key or hardcoding it.
[SAFE]: No obfuscation, persistence mechanisms, or unauthorized privilege escalation attempts were detected in the skill instructions or associated scripts.

nanobanana