The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process data from external sources, which creates a vulnerability to indirect prompt injection if the source data contains malicious instructions targeting the AI agent.
Ingestion points: The ETLPipeline class in scripts/run_etl_pipeline.py fetches data from external locations via _extract_api (HTTP GET requests) and _extract_csv.
Boundary markers: There are no explicit delimiters or instructions defined to separate external data from system instructions or to ignore embedded prompts within the data.
Capability inventory: The skill possesses the capability to write processed data to sensitive destinations, including SQL databases, AWS S3 buckets (via boto3), and the local file system.
Sanitization: Data is handled using standard libraries (Pandas), but there is no specialized logic to sanitize content against prompt injection or adversarial instructions.
[EXTERNAL_DOWNLOADS]: The skill utilizes network-capable libraries to download data from remote endpoints.
Evidence: scripts/run_etl_pipeline.py imports and uses the requests library to fetch JSON data from user-defined URLs and uses boto3 for interactions with AWS cloud storage services.

data-engineer