inventory-via-cypher
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill generates Cypher queries at runtime by interpolating user-provided inputs into predefined templates. Although the instructions include basic sanitization (escaping backslashes and single quotes), this pattern of code assembly from untrusted input represents a dynamic execution risk.
- [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Cypher injection) because it takes untrusted user input and places it directly into a query executed by the subimageRunCypher tool.
- Ingestion points: User-supplied resource types, product names, and filter values (e.g., name, region, state) are used to construct the query.
- Boundary markers: No formal delimiters or security boundaries are used; the skill relies on manual character escaping.
- Capability inventory: The skill uses subimageRunCypher (read-only) to execute the generated code.
- Sanitization: The skill instructs the agent to escape backslashes and single quotes before inlining literals, but this is insufficient to prevent all forms of injection.
Audit Metadata