flow-nexus-swarm
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation guides the user to download and install the 'flow-nexus' and 'claude-flow' packages from the NPM registry. These packages are essential components of the orchestration framework provided by the author (ruvnet).- [COMMAND_EXECUTION]: Users are instructed to perform system configuration tasks using standard shell commands like 'npm install', 'npx', and 'claude mcp add'. these actions are required to register the MCP server and initialize the workflow environment.- [PROMPT_INJECTION]: The skill architecture facilitates the processing of external data within workflows, identifying an indirect prompt injection surface.
- Ingestion points: Data enters the workflow context via the 'input_data' parameter in the 'mcp__flow-nexus__workflow_execute' function.
- Boundary markers: The current documentation does not describe the use of markers to separate instructions from processed data.
- Capability inventory: The skill has the ability to spawn agents, orchestrate complex tasks, and manage cloud-based execution streams.
- Sanitization: Specific input validation or sanitization routines are not detailed in the provided skill file.
Audit Metadata