The Agent Skills Directory

[COMMAND_EXECUTION]: The orchestration script scripts/pipeline_orchestrator.py dynamically compiles generated pipeline code using the compile() function for validation purposes. The compiled code is built from string templates that interpolate user-provided parameters such as pipeline names and shell commands, creating a potential vector for code injection if these inputs are derived from untrusted sources.
[DATA_EXFILTRATION]: The skill's analytical tools establish a surface for indirect prompt injection by processing external data files.
Ingestion points: scripts/data_quality_validator.py (reading CSV/JSON), scripts/etl_performance_optimizer.py (reading SQL queries), and scripts/pipeline_orchestrator.py (reading configuration files).
Boundary markers: The scripts do not employ delimiters or specific instructions to isolate data content from the agent's control instructions.
Capability inventory: The skill can read local filesystem data and generate executable orchestration scripts, which could be exploited in a multi-step attack.
Sanitization: Data handling focuses on schema and statistical validation without filtering for potential prompt injection patterns.
[SAFE]: The skill provides high-quality educational content and reference guides for professional data engineering. All external URLs and technology references pertain to established industry tools and services, such as Snowflake, BigQuery, Airflow, and dbt, and are documented neutrally.

senior-data-engineer