The Agent Skills Directory

[SAFE]: The skill follows security best practices by explicitly instructing the agent to ask for credentials and tokens rather than assuming defaults or searching for them in the environment.
[SAFE]: The instructions mandate a human-in-the-loop approach for sensitive actions, requiring the agent to summarize intended changes (deletions, metadata rewrites, or permission updates) before execution.
[SAFE]: No obfuscation, hardcoded credentials, or suspicious remote code downloads were found in the skill metadata or instructions.
[PROMPT_INJECTION]: The skill exhibits a standard indirect prompt injection attack surface which is typical for tools interacting with external data:
Ingestion points: Processes untrusted data from server logs, Docker inspection, and Jellyfin HTTP API responses.
Boundary markers: None explicitly defined in the instructions.
Capability inventory: Perform HTTP API writes, read system logs, and inspect container status.
Sanitization: No specific data sanitization or filtering is described.
Note: The overall risk is low and appropriate for the skill's management purpose.

jellyfin