The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or data exfiltration attempts were detected. The skill focuses on legitimate technical troubleshooting for a specific software tool.
[COMMAND_EXECUTION]: The skill uses cmux CLI commands to list, open, and close UI surfaces within the agent's environment. These operations are limited to the scope of the cmux tool and include instructions to anchor commands to the current $CMUX_WORKSPACE_ID and request user approval for significant changes.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to open and render external markdown files. While this represents a data ingestion surface, the risk is minimal as the primary goal is UI rendering rather than executing instructions contained within the files. The skill includes a 'Hard Rule' to verify rendering success rather than assuming success, which acts as a minor safeguard.

markdown-rendering