netboxlabs-platform-mcp
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill's primary function, 'Code Mode', relies on the agent generating Python code and executing it via the
netbox_executeandnetbox_search_schematools. Although the instructions describe a restrictive sandbox (blocking imports, dunder methods, and specific builtins), the core workflow is based on runtime code generation and execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from an external NetBox instance that could contain attacker-controlled content in fields like device names, descriptions, or custom fields.
- Ingestion points: Untrusted data enters the agent context through helper functions such as
get(),get_by_id(), andsearch()as documented inreferences/code-mode-helpers.md. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' markers when the agent processes retrieved data.
- Capability inventory: The skill enables the agent to execute code, perform write operations (create, update, delete), and interact with network services through the MCP platform.
- Sanitization: There are no explicit instructions for the agent to sanitize or validate data retrieved from NetBox before interpolating it into the Python scripts passed to the execution tools.
Audit Metadata