tilebox-datasets
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands using the 'tilebox' CLI, 'jq', and 'aws-cli' to manage Earth observation datasets, query metadata, and perform diagnostics. These operations often incorporate user-supplied input such as dataset slugs, collection names, and file paths. The skill also demonstrates writing local files (e.g., 'area.wkt') and generating code using 'tilebox dataset generate'.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources which could be manipulated to influence agent behavior.
- Ingestion points: The skill reads external geospatial files (GeoJSON, WKT), STAC metadata catalogs, and remote data stores (Zarr, Icechunk) through various provider APIs (AWS, Copernicus, ASF, Earthmover).
- Boundary markers: The instructions do not specify the use of clear delimiters or 'ignore embedded instructions' warnings when processing or interpolating data from these external sources into command arguments.
- Capability inventory: The skill possesses the ability to execute shell commands, perform network requests to geospatial providers, and write files to the local filesystem.
- Sanitization: While the skill uses structured parsing for JSON and WKT data, it does not explicitly mention sanitizing text fields or metadata values before they are used in the agent's decision-making context or CLI invocations.
Audit Metadata