hermes-agent

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

• Third-party content exposure detected (high risk: 0.90). The skill documentation explicitly describes fetching and ingesting public third-party content—e.g., web_search/web_extract and browser_navigate in references/features/tools.md and the web backend settings in references/configuration/configuration.md, plus installing/reading community skills from registries like skills-sh, well-known URLs and GitHub in references/features/skills.md—which the agent is expected to read and can materially influence its tool use and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The docs include a one-line installer that directly fetches and executes remote code at runtime (curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash), which is a high-risk remote-execution dependency.