ClawHub

Agent Team Orchestration

v3.0.0

Build and run multi-agent content production teams on OpenClaw with single-repo architecture, symlink-based file sharing, role-specialized AGENTS.md, and aut...

⭐ 1· 144·0 current·1 all-time
bysimonlin@simonlin1212
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name/description (multi-agent orchestration) match the included materials: SKILL.md, architecture docs, role templates, workflow, and a small setup script that creates workspaces and symlinks. There are no unrelated env vars, binaries, or external services requested.
ℹ
Instruction Scope
Instructions operate on local config and filesystem (creating ~/.openclaw workspaces, symlinks, and editing openclaw.json). This is appropriate for the purpose, but the docs explicitly recommend putting architecture/roster data into AGENTS.md so it is auto-loaded — that increases the amount of sensitive operational state injected into agent contexts and should be reviewed by the user.
✓
Install Mechanism
No install spec and no network downloads; the only executable content is a small bash script that creates directories, symlinks, and TOOLS.md files. Low-risk and proportionate to the skill's purpose.
✓
Credentials
No environment variables or credentials are requested. The script and instructions do write to ~/.openclaw and require filesystem permissions, which is expected for workspace setup. Ensure no secrets are placed into AGENTS.md, OUTPUT/, or KNOWLEDGE/.
ℹ
Persistence & Privilege
The skill does not request always:true or escalate privileges. However, it instructs the user to edit openclaw.json to grant the orchestrator 'subagents.allowAgents' — a necessary step for spawning but one that expands the main agent's abilities. Users should understand and trust the main agent before granting spawn permissions.
Assessment
This package appears internally consistent and implements what it claims: a local, symlink-based multi-agent pipeline. Before using it: (1) read the setup script (scripts/setup-team.sh) to confirm paths and behavior (it only creates dirs, symlinks, and TOOLS.md); (2) back up ~./openclaw/openclaw.json before modifying it and only add trusted agent IDs to subagents.allowAgents; (3) avoid placing secrets or credentials into AGENTS.md, OUTPUT/, or KNOWLEDGE/ (the docs recommend storing architecture info in AGENTS.md so it is auto-loaded — that increases exposure to any agent session that loads it); (4) consider filesystem permissions for the created workspaces so only intended agents/processes can read/write them; (5) run the script in a test environment first to verify behavior. If you want a deeper review, provide your openclaw.json and any AGENTS.md templates you plan to use (redact secrets) so I can check for accidental leakage of sensitive data.

Like a lobster shell, security has layers — review code before you run it.

content-productionvk97evtw57s0y38cq887a07p7gd83xgwylatestvk97evtw57s0y38cq887a07p7gd83xgwymulti-agentvk97evtw57s0y38cq887a07p7gd83xgwyorchestrationvk97evtw57s0y38cq887a07p7gd83xgwyworkflowvk97evtw57s0y38cq887a07p7gd83xgwy

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments