ROCm vLLM Deployment

Production-ready vLLM deployment on AMD ROCm GPUs. Combines environment auto-check, model parameter detection, Docker Compose deployment, health verification...

This skill appears to be a deployment helper and is not obviously malicious, but exercise caution before running it on a sensitive system. Specific actions to consider: 1) Inspect scripts locally (you already have them) before executing. 2) Avoid putting long-lived HF tokens in shell rc files; prefer ephemeral environment variables or a protected .env file with strict permissions. 3) Be aware check-env.sh will source ~/.bashrc (it will execute code from your rc file) and may create an empty ~/.bashrc if missing — run it in a controlled shell first or review/sanitize your rc file. 4) The scripts log the HF_TOKEN prefix and include status in reports stored under $HOME/vllm-compose — treat those files as sensitive. 5) If you want stronger assurance, ask the publisher for a full deployment workflow (the skill includes env-check and report generation but no deployment steps in these scripts) and confirm there are no external network endpoints or installers. If you plan to use this in production, prefer running on an isolated machine or container and store tokens in a secrets manager rather than in shell rc files.

Type: OpenClaw Skill Name: rocm-vllm-deployment Version: 1.0.0 The skill is classified as suspicious due to two main vulnerabilities. First, `scripts/check-env.sh` sources `~/.bashrc`, which could lead to arbitrary code execution if the user's `.bashrc` file is already compromised. Second, `scripts/generate-report.sh` embeds user-controlled parameters (e.g., `model-id`, `port`) directly into shell commands within the generated `DEPLOYMENT_REPORT.md`. If an AI agent or user were to blindly execute these commands from the report with specially crafted input, it could lead to command injection. While the skill itself does not exhibit clear malicious intent like data exfiltration or unauthorized persistence, these risky capabilities warrant a 'suspicious' classification.