The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts and toolchain binaries to automate neural network auditing tasks. All shell interactions use list-based arguments through subprocess.run, which effectively prevents command injection vulnerabilities.
[DYNAMIC_EXECUTION]: The generate_author_issue_bundle.py script uses importlib.import_module to verify environment dependencies. This functionality is restricted to a hardcoded list of known libraries (torch, onnx, linger) and does not accept arbitrary user input, ensuring secure execution.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves processing external data such as ONNX models and deployment logs. The system implements robust sanitization using shlex.quote when generating reproduction scripts from this data, mitigating risks of indirect prompt injection that might be embedded in model metadata or log output.
[SAFE]: No malicious patterns, data exfiltration to external domains, or unauthorized privilege escalations were detected. The skill's behaviors are consistent with its stated purpose of assisting in neural network design and toolchain integration for developer workflows.

lnn-model-design