The Agent Skills Directory

[SAFE]: No evidence of malicious behavior, persistence mechanisms, or credential harvesting was found. The skill aligns perfectly with its stated purpose of data retrieval.
[EXTERNAL_DOWNLOADS]: The skill fetches economic statistics from apis.datos.gob.ar. This is the official and well-known government API for public data in Argentina.
[COMMAND_EXECUTION]: Script execution is limited to standard HTTP requests using the requests library. No dangerous functions like eval(), exec(), or subprocess.run() are used.
[PROMPT_INJECTION]: Evaluated for potential indirect prompt injection vulnerability surface due to external data consumption.
Ingestion points: Data is fetched in scripts/fetch_indec.py from an external government API.
Boundary markers: Absent; the fetched data is interpolated into the agent context without explicit delimiters.
Capability inventory: The skill only performs HTTP GET requests and structured data processing; it lacks dangerous system capabilities.
Sanitization: Input parameters are validated by argparse and API responses are parsed as JSON and normalized.

indec