发现优质的 AI Agent 技能

- You want a **deterministic**, repeatable orchestration loop for OpenClaw (heartbeat + cron poll + subagents). - You want **evidence-gated** completion (no DONE without artifact paths). - You want a **lean portfolio queue** (`ACTIVITIES.md`) with external sprint-plan import via `Plan Path`.

Multi-agent orchestration for OpenClaw. Implements 5 proven patterns for coordinating multiple AI agents: Work Crew, Supervisor, Pipeline, Expert Council, and Auto-Routing.

This skill orchestrates sub-agents using user-provided task input. Core risks: - Prompt injection attempts embedded in task input - Unsafe task propagation to spawned sessions - Trust on external `openclaw` CLI behavior

How to publish this skill to ClawHub and monetization roadmap.

Multi-agent orchestration for OpenClaw. Implements 5 proven patterns for coordinating multiple AI agents: Work Crew, Supervisor, Pipeline, Expert Council, and Auto-Routing.

This skill orchestrates sub-agents using user-provided task input. Core risks: - Prompt injection attempts embedded in task input - Unsafe task propagation to spawned sessions - Trust on external `openclaw` CLI behavior

How to publish this skill to ClawHub and monetization roadmap.

Multipl is a job marketplace for AI agents.

来自 openclaw/skills 技能

This skill wraps the agent-browser CLI. Determine where the problem lies before reporting issues.

Thank you for your interest in contributing! This project aims to democratize high-level brand strategy using AI agents.

这个 skill 教你如何做一个更主动的 agent：不只是回答问题，而是预判需求、主动汇报、提前行动。