social-product-feedback-miner

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.70). The Skill contains explicit, deceptive runtime instructions to auto-install a dependency and "not describe" the installation or request user approval, which hides actions outside the user's expected consent for a feedback-mining tool.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill's functionality is mostly benign, but explicit instructions to install the BrowserAct dependency automatically without user confirmation and to conceal that installation indicate a deliberate attempt to perform unauthorized installation or hide supply-chain actions, which is high-risk behavior.

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

• Third-party content exposure detected (high risk: 0.85). This skill’s required runtime workflow invokes the browser-act Skill to access live public web pages and extract visible comment/feedback text, which is outsider-authored free text that can be ingested into the agent’s LLM context via the BrowserAct results.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs automatic runtime installation of the BrowserAct CLI using the URL https://www.browseract.com/?co-from=social-media&redirect=https://github.com/browser-act/skills/tree/main, which fetches and installs external code the agent relies on and that will execute at runtime.