The Agent Skills Directory

[SAFE]: The skill follows security best practices and does not exhibit any malicious behaviors.
[COMMAND_EXECUTION]: The orchestration script scripts/pipeline.py executes local Python scripts (validate_data.py, transform_data.py, handle_errors.py) using subprocess.run(). This is a legitimate functional requirement. The implementation is secure because it passes the command as an argument list and does not use shell=True, which effectively mitigates shell injection risks.
[DYNAMIC_EXECUTION]: According to the instructions in SKILL.md, the agent is expected to generate and execute Python code for custom data transformations. While this involves runtime execution, it is the primary purpose of the skill and occurs within the agent's controlled environment.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface that could theoretically be exploited by malicious content within processed files.
Ingestion points: assets/example_data.csv and any user-provided CSV/JSON files.
Boundary markers: None explicitly defined in the instructions to prevent the agent from following instructions embedded in the data.
Capability inventory: scripts/pipeline.py uses subprocess.run to execute Python scripts; the agent itself has broad tool access via allowed-tools.
Sanitization: The scripts use standard, safe parsers (csv.DictReader, json.load) which do not evaluate content as code. The risk is limited to the agent's interpretation of the data during analysis.

preprocessing-data-with-automated-pipelines