Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Homelab Cluster Management
v1.0.0Manage multi-tier AI inference clusters for homelabs. Health monitoring, expert MoE routing, automatic node recovery, and model deployment across Ollama and llama.cpp nodes. Covers GPU memory planning, Docker volume strategies for large models, sequential startup patterns to avoid CUDA deadlocks, and unified API gateways via LiteLLM.
⭐ 2· 876·4 current·4 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
The SKILL.md content (health checks, routing, Docker advice, SSH/RDP recovery, LiteLLM config) matches the stated 'Homelab Cluster Management' purpose. However, the skill declares no required binaries or environment variables while its instructions explicitly use docker, curl, ssh, RDP, and external vaults — a mild inconsistency: the runtime expects system/network tools and credentials even though none are listed in metadata.
Instruction Scope
The instructions remain within cluster-management scope: endpoint health checks, model routing logic, Docker volume strategies, sequential container startups, GPU memory planning, and recovery procedures. They instruct connecting to remote hosts (SSH/RDP) and operating Docker and HTTP endpoints, which is expected for this purpose. There is no obvious instruction to collect unrelated system data or to exfiltrate secrets, though the agent will be asked to handle sensitive credentials if used.
Install Mechanism
This is an instruction-only skill with no install spec and no code files, which is low-risk from an installation perspective (nothing is written to disk by the skill package itself).
Credentials
The skill requests no declared environment variables or credentials, but the guidance assumes use of SSH/RDP credentials and external vaults (Azure/HashiCorp) and references API keys in LiteLLM config snippets. It's reasonable for a management skill to require such credentials at runtime, but the metadata does not document them — users should not supply secrets implicitly without clear prompts and should prefer a vault-backed workflow as the doc suggests.
Persistence & Privilege
always:false (default) and autonomous invocation enabled (also default). The skill does not request permanent 'always' presence or attempt to modify other skills/config. No persistence or escalation behaviors are declared.
Assessment
This skill appears to be coherent for homelab cluster management, but it expects the agent (or the operator) to run network and system commands (docker, curl, ssh, RDP) and to supply remote credentials or vault access at runtime even though none are declared. Before installing or enabling it:
- Verify provenance (source is 'unknown' and there's no homepage). Consider running in an isolated test environment first.
- Do not provide long-lived credentials directly to the skill; use a secrets vault as recommended and prefer short-lived credentials.
- Require explicit human confirmation before the agent performs SSH/RDP, restarts Docker, or copies model files — those actions can be disruptive.
- Audit the full, untruncated SKILL.md to confirm there are no instructions that run arbitrary downloaded code or call unknown external endpoints.
- If you allow autonomous runs, restrict the agent's network and credential scope (least privilege) and log all actions so you can review recovery operations and container restarts.
If you want, provide the full SKILL.md (it was truncated in the package) and any provenance or author contact so I can re-check for missing or risky instructions.Like a lobster shell, security has layers — review code before you run it.
latestvk97exnp9jkbzhj51h511cghmnn8119ce
License
MIT-0
Free to use, modify, and redistribute. No attribution required.