The Agent Skills Directory

[Data Ingestion Surface]: The skill includes patterns for bulk-importing CSV files and querying Dataverse tables. This creates an interface where the agent processes external data, which is a potential consideration for indirect prompt injection if the source content is untrusted.
Ingestion points: data/tickets.csv and Dataverse record queries.
Boundary markers: Not explicitly defined in the provided script examples.
Capability inventory: Python script execution, file system read access, and Dataverse record creation/update/deletion.
Sanitization: Basic field mapping is demonstrated, though complex sanitization logic is left to the implementation.
[External Dependencies]: The skill recommends installing official packages from a well-known service and trusted organization. This is a standard practice for enabling SDK functionality and supporting data analysis with tools like pandas.
[Dynamic Path Management]: The script examples use sys.path.insert to reference local utility scripts. This is a common pattern in Python development for organizing project-specific logic and does not involve untrusted path construction.
[Credential Management]: The documentation provides guidance on using environment variables and .env files for managing secrets. This is an industry-standard practice for avoiding hardcoded credentials in source code.

dv-python-sdk