validate-data
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data files and user inputs, creating an attack surface where malicious instructions embedded in that data could influence the agent's actions.\n
- Ingestion points: Workflow step 1 in
SKILL.mdrequires the agent to inspect "available files, runtime, versions, inputs."\n - Boundary markers: The skill instructions do not define specific delimiters or provide guidance to ignore embedded instructions within the data being processed.\n
- Capability inventory: The skill can perform data transformations and has network capabilities via the
sandbase_runtool to interact with external datasets and models.\n - Sanitization: No explicit procedures for sanitizing, escaping, or filtering content from external sources are mentioned in the workflow.\n- [EXTERNAL_DOWNLOADS]: The skill utilizes external APIs for data discovery, inspection, and execution tasks.\n
- Evidence:
references/sandbase-api-map.mdlists tools such assandbase_runandsandbase_discoverfor interacting with external services.\n - Analysis: These are vendor-provided tools intended for data validation involving external datasets. The skill includes safeguards such as requiring user confirmation before performing paid operations.
Audit Metadata