The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the runcomfy CLI tool to perform image editing by routing intents to specific models. It provides templates for shell commands that incorporate user-provided prompts and URLs.
[EXTERNAL_DOWNLOADS]: Recommends the installation of the @runcomfy/cli package from npm, which is the official CLI for the RunComfy platform.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources which could contain malicious instructions.
Ingestion points: The skill accepts untrusted data through the image_urls, images, image, and mask_image fields in SKILL.md.
Boundary markers: Data is structured in JSON format for the CLI, but there are no explicit delimiters or instructions to the agent to ignore content embedded in the images or prompt strings.
Capability inventory: The skill performs shell command execution through the runcomfy CLI as described in SKILL.md.
Sanitization: Escaping of single quotes in shell commands is illustrated in the documentation, but active sanitization of the remote content is managed by the vendor's platform and CLI.

image-edit