hackerone

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md and README explicitly instruct the agent to "fetch program data and guidelines" and "download scope CSV" from a HackerOne program URL (and to parse the CSV including an instruction column), which means it ingests open third-party program pages and user-provided CSV content that directly guide testing behavior and agent actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly fetches HackerOne program pages (e.g., https://hackerone.com/program-name) at runtime and uses the downloaded program guidelines and scope CSV to drive pentester agents' instructions, so remote content can directly control agent behavior.