The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/pipeline_orchestrator.py dynamically generates code for Airflow, Prefect, and Dagster workflows and uses the compile() function for syntax validation. Evidence: Lines 270, 335, and 396 in scripts/pipeline_orchestrator.py call compile(code, '<string>', 'exec'). This pattern of assembling code from user-defined configuration and parsing it at runtime is a security risk if the input values are not strictly sanitized.
[INDIRECT_PROMPT_INJECTION]: The skill provides tools to ingest and analyze external data files, creating a potential surface for prompt injection attacks. Evidence Chain: 1. Ingestion points: scripts/data_quality_validator.py and scripts/etl_performance_optimizer.py load data from CSV, JSON, and SQL files. 2. Boundary markers: Absent; the scripts do not use specific delimiters or instructions to the agent to disregard embedded directives. 3. Capability inventory: The skill possesses code generation and validation capabilities through pipeline_orchestrator.py. 4. Sanitization: Absent; while the scripts perform structural validation, they do not filter data for semantic instructions designed to influence the agent's behavior.

senior-data-engineer