The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes an external binary RVT2IFCconverter.exe via subprocess.run to perform file conversions. This is the primary function of the skill and is explicitly described in the documentation.
[INDIRECT_PROMPT_INJECTION]: The validate_output function in SKILL.md reads up to 1000 characters from generated or provided IFC files to check for valid headers and version strings. This processing of untrusted file data into the agent's context creates a potential surface for indirect prompt injection.
Ingestion points: Reading content from .ifc files in SKILL.md via f.read(1000).
Boundary markers: No delimiters or safety instructions are used when reading the file content.
Capability inventory: The skill uses subprocess.run for command execution and has filesystem permissions for reading and writing files.
Sanitization: The skill only performs basic substring checks for version identification and does not sanitize or escape the content read from the file.

rvt-to-ifc