← Back to Skills Marketplace
51
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install yihui-agent-swarm
Description
Multi-agent orchestration framework for OpenClaw. Define roles, route tasks, manage state, and coordinate agent teams using structured YAML configs and a pro...
Usage Guidance
This skill appears to be a legitimate multi-agent orchestration design, but it omits practical deployment details that matter for security. Before installing or enabling it: 1) Ask the publisher which runtime connectors are required and what credentials (Feishu API token, MemPalace/MemClaw endpoints and keys) you must provide — do not supply broad tokens without least-privilege scopes. 2) Confirm where memories are stored, retention/erasure policies, and whether PII or secrets will be redacted before writing to MemPalace/Cortex. 3) Test in a restricted sandbox account (with synthetic data and limited credentials) first. 4) Ensure a Critic role or policy enforces privacy/security checks and that memory-writing is auditable. If the publisher cannot explain why no env/credential requirements are declared, treat this mismatch as a red flag and avoid giving production credentials.
Capability Analysis
Type: OpenClaw Skill
Name: yihui-agent-swarm
Version: 1.0.0
The bundle provides a comprehensive framework for multi-agent orchestration, defining roles (Master, Specialist, Critic, Router) and communication protocols using structured YAML. The logic is focused on task routing, memory management via MemPalace/Cortex, and quality assurance, with explicit security and privacy checks defined in 'configs/roles/critic.yaml'. While 'SKILL.md' and 'configs/team.yaml' include specific Feishu Open IDs (e.g., ou_7d0630bcf6e7fb2c35c2033be5e23722), these are clearly documented as examples from a real production setup (YiHui team) and are intended to be replaced by the user during configuration.
Capability Assessment
Purpose & Capability
The SKILL.md describes integrations with external services (MemPalace, Cortex/MemClaw, and Feishu channel targets) and requires structured message routing; however the registry metadata declares no required environment variables, credentials, or config paths. In practice those integrations typically need API tokens/endpoints. The absence of declared credentials or dependency hints is an incoherence between stated purpose and declared requirements.
Instruction Scope
The runtime instructions focus on routing, role semantics, and writing/recalling memory (dual recall, immediate writes to MemPalace/Cortex). They do not instruct reading arbitrary local files or executing shell commands (good), but they do mandate automatic memory writes and cross-agent state persistence without explicit guidance on handling sensitive or PII data or retention/expiration policies. That broad memory-write behavior could inadvertently persist sensitive information.
Install Mechanism
This is instruction-only with no install spec and no bundled code — low install-time risk. Nothing is downloaded or written by an installer as part of the skill package itself.
Credentials
Although the skill expects use of Feishu channel targets and memory services, it lists no required environment variables, no primary credential, and no required config paths. Real deployment would likely require Feishu API keys, MemPalace/MemClaw endpoints and credentials, or platform-tool bindings; those secrets are not declared here, creating an opaque mismatch about what credentials will actually be needed and where they must be provided.
Persistence & Privilege
always is false (no forced global inclusion). The skill is user-invocable and allows model invocation (the platform default). It does not request to modify other skills or system-wide configs in the SKILL.md. Autonomous invocation combined with the missing credential/deployment details increases blast radius, but autonomous invocation alone is expected.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install yihui-agent-swarm - After installation, invoke the skill by name or use
/yihui-agent-swarm - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release: multi-agent orchestration based on real YiHui team setup
Metadata
Frequently Asked Questions
What is YiHui Agent Swarm?
Multi-agent orchestration framework for OpenClaw. Define roles, route tasks, manage state, and coordinate agent teams using structured YAML configs and a pro... It is an AI Agent Skill for Claude Code / OpenClaw, with 51 downloads so far.
How do I install YiHui Agent Swarm?
Run "/install yihui-agent-swarm" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is YiHui Agent Swarm free?
Yes, YiHui Agent Swarm is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does YiHui Agent Swarm support?
YiHui Agent Swarm is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created YiHui Agent Swarm?
It is built and maintained by 辉哥 (@1yihui); the current version is v1.0.0.
More Skills