multi-agent-orchestrator

This skill enables one-click generation of multiple AI agents based on a user prompt, outputs their organizational structure, and visualizes their collaborat...

This skill appears to do what it says (generate agent descriptions and output Mermaid diagrams) but exercise caution before installing/running: 1) The scripts call an external binary named 'manus-render-diagram' via subprocess.run but the skill does not declare or provide an install/source for that tool — verify where you would get that binary and install it from a trusted source before running. 2) The examples use absolute home paths and the scripts create directories and write files; run the skill in a safe/sandboxed directory to avoid unexpected file writes. 3) Review the scripts locally (they are short and readable) and, if you plan to run them on sensitive systems, ensure they are executed with least privilege (not as root). 4) If you need to allow the agent to run this skill autonomously, be aware the skill can invoke local binaries — only enable autonomous invocation if you trust the runtime environment and any external tools it calls. If you want higher assurance, ask the author to declare required binaries and provide installation instructions or a verified source for 'manus-render-diagram'.

Type: OpenClaw Skill Name: multi-agent-orchestrator Version: 1.0.0 The skill is designed to generate and visualize multi-agent organizational structures based on user prompts. It uses subprocess.run safely with list arguments to execute a local diagram rendering tool (manus-render-diagram) and manage output directories. The logic in scripts/orchestrate_and_visualize.py and scripts/generate_agents.py is transparent, aligns with the stated purpose in SKILL.md, and contains no evidence of data exfiltration, backdoors, or malicious intent.