The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions designed to bypass safety filters, extract system prompts, or override core agent behavior were detected. Content is restricted to brand management guidance.
[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file path references (e.g., .ssh, .env), or network exfiltration patterns (curl, wget) were found.
[Remote Code Execution] (SAFE): The skill does not download or execute remote scripts. No shell pipes or dynamic code execution patterns are present.
[Indirect Prompt Injection] (SAFE): While the skill prompts for user-provided brand data, it lacks any high-privilege capabilities (like file writing or subprocess execution) that could be exploited via malicious input. Standard AI safety guardrails apply.
[Persistence & Privilege Escalation] (SAFE): No commands were found that attempt to modify system configuration, startup scripts, or gain administrative privileges.

brand-management