hive-setup

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

• Third-party content exposure detected (high risk: 0.90). This skill clones tasks from a Hive server using "hive task clone " and explicitly runs/inspects repository-provided files (prepare.sh, eval/eval.sh, program.md), which are untrusted third-party content the agent reads and executes as part of the setup workflow, enabling indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill includes a runtime installation command that fetches and pipes a remote shell script to sh ("curl -LsSf https://astral.sh/uv/install.sh | sh"), which executes remote code from https://astral.sh and thus poses a direct execution risk.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill instructs installing system packages with sudo, running remote install scripts (curl | sh), and executing cloned task prepare.sh (arbitrary code), all of which modify system state and can compromise the host.