Multi Agent Dev Team

Set up and manage 2–10 agent dev teams with customizable roles, workflows, multiple parallel teams, and auto-assigned AI models for collaborative software de...

This package is internally consistent for setting up multi-agent teams, but take a few precautions before running it: 1) Backup your OpenClaw config (~/.openclaw/openclaw.json) as the docs instruct. 2) Review the exact openclaw.json changes the wizard will make (especially merging allowAgents and any subagent entries) to avoid unintentionally exposing or overwriting agent lists. 3) Inspect the default model ID strings the wizard proposes (some defaults reference a 'custom-llmapi-lovbrowser-com' prefix); confirm they map to providers you trust or override them with your registered model IDs. 4) Run the wizard in a test environment first (non-production) to validate behavior. 5) Note the package author/homepage are not provided — treat it like third-party community code and review the code before granting access to production configs.

Type: OpenClaw Skill Name: multi-agent-team-by-halfmoon82 Version: 2.2.2 The skill bundle provides a multi-agent management framework but includes high-risk instructions in SKILL.md that direct the AI to autonomously modify its core configuration (openclaw.json) and establish persistence via 'OpenClaw Cron' jobs. Specifically, the 'Standard Post-Setup Workflow' mandates that the agent rewrite its own skill mappings and security allowlists (allowAgents) weekly. While no explicit data exfiltration or external malicious endpoints were identified, this self-modifying behavior and the requirement for broad read/write/exec permissions in clawhub.yaml create a significant attack surface for privilege escalation and persistent state alteration.