The Agent Skills Directory

[Prompt Injection] (SAFE): No malicious instructions, override attempts, or bypass markers were detected. The skill uses 'Tell AI' blocks as helpful templates for the user to interact with the agent in a structured manner.
[Data Exposure] (SAFE): No hardcoded credentials, sensitive file paths, or unauthorized network operations were found. The documentation correctly uses placeholders like [phone] and [DSN] for user-provided configuration.
[Remote Code Execution] (SAFE): The skill does not include any scripts or commands that download and execute external code. It recommends reputable services (e.g., Sentry, UptimeRobot) for manual setup by the user.
[Persistence & Privilege Escalation] (SAFE): No patterns related to system persistence, unauthorized privilege acquisition, or suspicious command-line execution were identified. The shell commands provided in the documentation are for manual user reference during rollbacks or git operations.
[Indirect Prompt Injection] (SAFE): While the skill suggests that users provide error logs to the AI for analysis, it does not include any automated ingestion mechanisms that would bypass user oversight.

monitor