The Agent Skills Directory

[COMMAND_EXECUTION]: Rule 3 in SKILL.md provides a shell command template (grep) designed to search through local Python stubs. This template uses a methodName placeholder that is intended to be filled by the agent. If the agent populates this placeholder using unsanitized data from a user (e.g., a string containing ;, |, or $()), it could lead to arbitrary command execution on the host system.
[PROMPT_INJECTION]: The skill's core workflow (Rule 2) involves the agent generating and deploying Python scripts to persistent local directories (~/Library/Application Support/Autodesk/Autodesk Fusion 360/API/Scripts/). This creates an attack surface where an indirect prompt injection—such as a malicious instruction hidden in a log file provided by a user—could trick the agent into writing and executing harmful code in a trusted application path.
[EXTERNAL_DOWNLOADS]: The section-index.md file contains numerous links to official Autodesk documentation (help.autodesk.com). These references are directed toward a well-known service related to the skill's primary purpose and do not represent a security risk.

fusion360-scripting