The Agent Skills Directory

[EXTERNAL_DOWNLOADS]:
The skill instructs the installation of standard data connectivity libraries (psycopg2-binary, ibm_db, ibm-cos-sdk, confluent-kafka) from the official PyPI registry.
Unverifiable Dependency: It explicitly instructs the installation of ibm-aspera-connect via pip, but this package name is not found on the official Python Package Index (PyPI). This presents a supply chain risk, as an attacker could register this name to distribute malicious code.
[COMMAND_EXECUTION]: Provides guidance for using the ibmcloud CLI and its associated plugins (cloud-databases, event-streams) as well as the aws CLI for managing cloud resources and transferring data.
[PROMPT_INJECTION]:
Ingestion points: The skill allows the agent to retrieve and process data from external PostgreSQL, MongoDB, Cloudant databases, and Cloud Object Storage buckets.
Boundary markers: None of the provided code examples use delimiters or instructions to ignore potential commands embedded within the retrieved data.
Capability inventory: The skill has access to powerful tools including shell and write_file.
Sanitization: There is no implementation of input validation or sanitization for content fetched from external services.
[CREDENTIALS_UNSAFE]: The skill uses static example passwords (e.g., "StrongPass123!") in CLI commands and refers to local JSON files for storing service credentials, which can encourage insecure secret management.

ibm-data