The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches and installs the flow-nexus and claude-flow packages from the NPM registry as part of the environment setup.
[COMMAND_EXECUTION]: Provides instructions for the user to execute shell commands for global package installation, platform registration, account authentication, and configuring the MCP server within the local environment.
[INDIRECT_PROMPT_INJECTION]: The skill's architecture processes external data which could serve as a vector for indirect prompt injection attacks.
Ingestion points: Untrusted data enters the agent context through task descriptions in mcp__flow-nexus__task_orchestrate, workflow data in mcp__flow-nexus__workflow_execute, and information gathered by research agents via web searches.
Boundary markers: There are no explicit delimiters or instructions shown to help the agent distinguish between its core instructions and the external data being analyzed or processed.
Capability inventory: The skill has the ability to manage cloud swarms, coordinate multi-agent tasks, and retrieve lists of files generated during execution via mcp__flow-nexus__execution_files_list.
Sanitization: No validation or sanitization procedures are documented for the input data before it is handled by the swarm agents.

flow-nexus-swarm