flow-nexus-platform

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The tool mcp__flow-nexus__sandbox_execute allows the agent to run arbitrary code (JavaScript, Python) within created sandboxed environments.
  • [REMOTE_CODE_EXECUTION]: The mcp__flow-nexus__template_deploy tool allows for the deployment of application templates from the Flow Nexus platform, which involves executing code logic defined in those templates.
  • [INDIRECT_PROMPT_INJECTION]: The mcp__flow-nexus__seraphina_chat tool accepts user-provided chat history and messages which are processed by an AI assistant capable of tool execution.
  • Ingestion points: The message and conversation_history parameters in mcp__flow-nexus__seraphina_chat function in SKILL.md.
  • Boundary markers: No delimiters or "ignore embedded instructions" warnings are specified for the AI assistant input.
  • Capability inventory: The assistant can create swarms, orchestrate agents, and deploy code as described in the system utilities section of SKILL.md.
  • Sanitization: No explicit sanitization, validation, or filtering of external chat content is documented.
  • [DYNAMIC_EXECUTION]: mcp__flow-nexus__sandbox_execute and mcp__flow-nexus__template_deploy support runtime execution of code strings and template logic provided during the session.
  • [EXTERNAL_DOWNLOADS]: Sandbox management tools like mcp__flow-nexus__sandbox_create and mcp__flow-nexus__sandbox_configure allow for the installation of external packages from standard registries such as npm and pip.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 06:15 PM
Security Audit — agent-trust-hub — flow-nexus-platform