The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data via the $ARGUMENTS variable to define furniture design constraints.
Ingestion points: The skill uses $ARGUMENTS at the start of the design phase in SKILL.md to extract object type, dimensions, and materials.
Boundary markers: The input is clearly delimited by a TARGET header section.
Capability inventory: The skill is purely generative and lacks access to sensitive tools. It does not perform file writes, network operations, or subprocess executions.
Sanitization: While there is no explicit input sanitization, the restricted output format (design specifications, ASCII diagrams, and tables) mitigates the risk of malicious payload execution.
[SAFE]: No evidence of prompt injection, data exfiltration, obfuscation, or unauthorized command execution was found. The instructions are focused entirely on structural engineering and fabrication logic for sheet materials.

cnc-furniture