commercial-skills
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external business data, creating a vulnerability surface for instructions embedded within those files to manipulate the agent's logic.
- Ingestion points: Untrusted data enters the agent context through files such as RFP PDFs,
.csvpipeline exports, and.docxpartner agreements (SKILL.md). - Boundary markers: The provided instructions do not define explicit boundary markers or include warnings for the agent to ignore instructions embedded within the processed commercial artifacts.
- Capability inventory: The skill uses the ingested data to drive deterministic routing decisions, generate commercial digests, and cite pricing/deal desk benchmarks (SKILL.md).
- Sanitization: There are no described mechanisms for sanitizing, validating, or escaping content found within external documents before it is used in the prompt context.
Audit Metadata