The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently instructs the agent to execute python3 freecad-client.py using shell heredocs (<<'PY') to run Python snippets on the local system.
[REMOTE_CODE_EXECUTION]: The skill implements a client-server architecture (freecad-client.py and freecad-server.py) to execute Python code within the context of a live FreeCAD process. This allows for arbitrary Python execution, which typically includes access to powerful system libraries like os and subprocess.
[PROMPT_INJECTION]: The skill is highly susceptible to indirect prompt injection due to its core design of translating user modeling requests into executable Python scripts.
Ingestion points: Untrusted user requests for 3D modeling and CAD operations enter the agent's context and are used to generate code.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore or escape potentially malicious instructions embedded within user modeling requests.
Capability inventory: The agent has the ability to execute arbitrary Python code within FreeCAD and invoke shell commands via the freecad-client.py script.
Sanitization: The skill does not provide any mechanisms for sanitizing, validating, or sandboxing the generated Python code before it is executed.
[EXTERNAL_DOWNLOADS]: The skill references official FreeCAD documentation and download URLs (e.g., freecad.org, github.io). These are recognized as legitimate domain-specific resources and do not pose a security risk.

freecad-ai