hive-setup

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). This is a direct link to an install shell script on a third‑party domain (astral.sh); piping or running it would execute arbitrary code on your machine and is high risk unless you verify the script's contents and trust the publisher.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and clones tasks from hive servers (Step 3/4: hive task list and hive task clone <task-id>), accepts custom server URLs (Step 2), and runs repository-supplied files like prepare.sh and eval scripts (Step 5), meaning it ingests and executes untrusted third‑party content from public repos/servers that can materially influence agent behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill can execute remote code at runtime — it explicitly runs "curl -LsSf https://astral.sh/uv/install.sh | sh" to install uv if needed and also clones task repositories via "hive task clone" (from the server URL) and runs their prepare.sh, so fetched external content is executed and can control agent behavior.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). Flagged because the skill instructs the agent to install system packages using sudo (e.g., sudo apt-get install), run global installers (pip/pipx/uv, curl | sh), and execute task-provided prepare.sh — actions that modify the host system state and can require or obtain elevated privileges.