carto-site-selection
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines workflow generation patterns that ingest untrusted user-supplied data, such as database table names, filtering thresholds, and scoring weights, which are then interpolated into functional workflow nodes.\n
- Ingestion points: User-specified table names in
native.gettablebynameandnative.saveastable, filtering predicates innative.where, and numeric weights innative.selectexpression.\n - Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings for the data being processed.\n
- Capability inventory: The skill utilizes tools capable of reading from (
native.gettablebyname) and writing to (native.saveastable) the connected database environment.\n - Sanitization: No validation or escaping of user-provided strings is described in the workflow construction steps.
Audit Metadata