The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the claude-flow package from the NPM registry to enable swarm orchestration. This is a primary component of the framework described by the author.
[COMMAND_EXECUTION]: Orchestration workflows use tools such as mcp__claude-flow__parallel_execute and mcp__claude-flow__task_orchestrate to run specific CLI commands like npm run test or research search queries.
[DYNAMIC_EXECUTION]: The skill documents how to assemble and trigger complex logic at runtime through tools like mcp__claude-flow__workflow_create and mcp__claude-flow__parallel_execute, which accept command and task strings for execution.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The patterns process data from external sources such as web searches, academic papers, and codebases during research and development phases.
Boundary markers: No explicit delimiters are shown in the pattern examples to separate external data from coordination instructions.
Capability inventory: The swarm has access to shell execution, network operations for research, and file system writes for memory backups.
Sanitization: Sanitization of ingested data is not explicitly detailed in the provided orchestration patterns.

swarm-advanced