navan-deploy-integration
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes instructions to fetch and process data from the Navan API (e.g., expenses, users). This creates a potential surface for indirect prompt injection if the external data contains malicious instructions meant to be interpreted by the agent.
- Ingestion points: Data is ingested from
https://api.navan.com/v1/expensesandhttps://api.navan.com/v1/usersas documented inSKILL.md. - Boundary markers: There are no explicit instructions or delimiters to warn the agent against executing commands embedded within the fetched data.
- Capability inventory: The skill utilizes
Bash,Write, andEditcapabilities which could be targeted by an injection attack. - Sanitization: The provided scripts use
jqfor JSON parsing, which validates the structure of the data but does not sanitize the content for natural language instructions.
Audit Metadata