tableau-next-asset-create
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides functionality for ingesting and profiling external data files (CSV/Excel). This creates an attack surface where malicious content within those files could attempt to influence the agent's behavior.
- Ingestion points: Data enters the context through tools like
run_query,run_semantic_query, andinfer_object_schema(documented inreferences/tasks/ingest-flat-file.mdandreferences/large-flat-file-handling.md). - Boundary markers: The skill uses explicit profiling and classification steps (Step 0.5 in
references/build-workflow.md) to separate business fields from system fields. - Capability inventory: The skill can execute SQL queries, create/delete assets, and manage user shares/permissions across various scripts.
- Sanitization: The skill contains a strict 'infer-don't-read' protocol (documented in
references/large-flat-file-handling.md) which explicitly forbids reading the entire file into the context window, effectively mitigating the risk of large-scale prompt injection from data files.
Audit Metadata