The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external image URLs, which represents a potential attack surface for indirect prompt injection where malicious content in an image could attempt to influence agent behavior. * Ingestion points: The image and image_urls parameters within the JSON payload in SKILL.md. * Boundary markers: The skill instructions recommend using JSON structure and only processing URLs explicitly provided by the user. * Capability inventory: Shell access is restricted to the runcomfy CLI via the Bash tool. * Sanitization: Guidance provided in the skill documentation suggests monitoring for unexpected model behavior and verifying reference assets.
[EXTERNAL_DOWNLOADS]: The skill includes instructions to install the @runcomfy/cli package via npm. This package is the official tool provided by the skill author's organization.
[COMMAND_EXECUTION]: The skill uses the Bash tool to interact with the runcomfy CLI. This capability is appropriately scoped in the skill frontmatter to limit execution only to that specific command.

relight