automation

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

• Third-party content exposure detected (high risk: 0.90). The skill's Plugin Preset and custom-automation workflow explicitly fetch and load plugins and tarballs from public GitHub repos, git URLs, and HTTPS/S3 URLs (see "Plugin Preset" in SKILL.md and "Tarball Path Formats"/"Plugin Source Formats" in references/custom-automation.md), meaning the agent ingests and executes untrusted third‑party code that can materially change its actions.

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

• Potentially malicious external URL detected (high risk: 1.00). Yes — the skill fetches and loads plugins at runtime from git sources (e.g., github:owner/repo or https://github.com/owner/repo.git) which inject skills/MCP/hooks/commands into the agent conversation (directly affecting prompts/behavior), and the custom automation setup explicitly executes remote code via curl -LsSf https://astral.sh/uv/install.sh | sh.