The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/run_sample_calibration.py utilizes subprocess.check_call to create a virtual environment and os.execv to re-execute itself within that environment. This mechanism is used to ensure the presence of the requests library without requiring manual installation by the user.\n- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill's Python script uses pip to install the requests library if it is missing. Additionally, the skill references the NVIDIA NGC registry for downloading supplemental datasets, which is an expected behavior for a vendor-provided tool.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that ingests data from local synthetic files including alignment_data.json, layout.png, and GT.zip.\n
Ingestion points: Files are read from the assets/ directory and uploaded to the microservice in scripts/run_sample_calibration.py.\n
Boundary markers: The instructions do not include specific delimiters or warnings to ignore potentially malicious content within these data files.\n
Capability inventory: The skill has the capability to execute shell commands (e.g., unzip, port scanning), write to the filesystem (extracting data), and perform network operations via REST API calls.\n
Sanitization: Content validation is handled by the target AMC microservice's API endpoints rather than the skill's scripts.

amc-run-sample-calibration