The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute specific local command-line utilities including tpacker and tvalidator for model optimization and consistency checks. It also performs shell commands to verify the environment, such as checking tool versions and using git to verify the state of the local repository.
[DYNAMIC_EXECUTION]: The skill utilizes cmake to compile target platform simulation libraries during the validation phase if they are not pre-built. It also executes dynamic Python code via python -c to inspect the versions of installed dependencies like linger and pythinker.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data in the form of quantised ONNX models and NPY input files. While these are binary formats, they serve as a primary ingestion surface. The skill mitigates risks by requiring an initial validation step using onnx.checker.check_model to verify graph integrity before proceeding with deployment tasks.

thinker-model-pack-validation