The Agent Skills Directory

[UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill depends on the @three-ws/mcp-server tool and communicates with the three.ws domain to perform 3D model generation and hosting. These resources are integral to the skill's stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided text descriptions to generate 3D assets, which introduces a surface for indirect prompt injection.
Ingestion points: The prompt parameter in the forge_free tool call defined in SKILL.md.
Boundary markers: Absent. The instructions do not specify any delimiters (such as triple quotes) or negative constraints to prevent the agent from following instructions embedded in the user prompt.
Capability inventory: The skill calls the forge_free tool to interact with a remote API and returns durable URLs (glbUrl, viewerUrl) for user consumption.
Sanitization: Absent. No logic is provided to escape, filter, or validate the contents of the user-supplied prompt before it is sent to the backend pipeline.

forge-3d