The Agent Skills Directory

[SAFE]: The skill follows security best practices by recommending the use of verified package managers (npm/npx) for tool installation and explicitly warning against unsafe methods like piping remote scripts into a shell.
[SAFE]: The skill implements secure credential management for the RunComfy CLI, documenting the use of environment variables and local configuration files with restricted permissions (0600) for storing API tokens.
[SAFE]: The skill correctly identifies the attack surface for indirect prompt injection, where untrusted third-party video content (e.g., embedded metadata) could influence agent behavior, and provides clear guidance for the agent to verify user intent and evaluate source reliability.

video-inpainting