The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted data from external URLs and local files. However, it incorporates robust security best practices to mitigate this risk, including mandatory schema validation (enforcing column names and data types) and data integrity checks (detecting gaps in time-series data).
Ingestion points: fetch_with_retry(url) and load_and_validate(path) methods.
Boundary markers: The skill explicitly implements an EXPECTED_SCHEMA validation layer and uses assert statements to verify data structure before use.
Capability inventory: Utilizes requests for network operations and polars for file system read access.
Sanitization: Implements schema enforcement, type validation, and logical consistency checks (gap detection) to prevent downstream processing of malicious or malformed data.
[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected. The skill uses standard libraries (pandas, polars, requests) and vendor-specific tools (ml4t-data) consistent with the author's identity and the stated purpose of financial data management.

ml4t-fetch-data