homepage-audit

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

• Third-party content exposure detected (high risk: 1.00). The skill's Context Loading Gates explicitly require "A live URL (fetch with web_fetch if available)" and instruct the agent to load and audit that public webpage, so it will fetch and interpret arbitrary third-party site content as part of its decision-making.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires fetching a user-supplied live homepage (using "web_fetch" at runtime) and injecting that page's content into the audit context, so an attacker-controlled page URL (the user-provided homepage URL fetched via web_fetch) could directly influence prompts/outputs.