awesome-hermes-agent-ecosystem

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to clone and use community GitHub repos and to use MCP github/filesystem servers (e.g., multiple git clone commands, "use github MCP to list open PRs", and "observe my next 3 git workflows and create a skill"), which causes the agent to fetch and read untrusted, user-generated content from public sites that can influence decisions and tool actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly runs remote installer code at runtime (curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash) and instructs cloning and running the hermes-agent repo (https://github.com/NousResearch/hermes-agent then npm install && npm start), which fetches and executes external code that directly controls the agent environment.