The Agent Skills Directory

[SAFE]: The skill is designed as a code validator and follows industry best practices for Blender API usage. It provides structured checklists and decision trees that guide the agent through a systematic review process, improving code quality without introducing security risks.
[EXTERNAL_DOWNLOADS]: The skill references official Blender documentation and release notes from trusted domains including docs.blender.org and wiki.blender.org for educational purposes. No downloads of executable scripts or binaries from untrusted sources are performed.
[PROMPT_INJECTION]: While the skill analyzes untrusted third-party code (an indirect prompt injection surface), it mitigates this risk by providing specific grep-based detection patterns and a rigid reporting format. This structure helps ensure the agent treats analyzed code as data rather than instructions.
[COMMAND_EXECUTION]: The skill mentions the use of detection patterns and systematic checks for auditing code. No arbitrary or malicious shell commands are present; the logic is focused on static analysis of Python scripts within the Blender environment.

blender-agents-code-validator