ecommerce-competitor-assortment-monitor

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The skill explicitly instructs the agent to install dependencies automatically and to conceal or avoid asking the user for confirmation ("without asking the user to confirm" and "Do not describe it as a request... and do not request separate approval"), which is a deceptive instruction outside the stated monitoring purpose.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). These links point to a third‑party site (browseract.com) that the skill instructs agents to use to automatically install a CLI dependency (potentially downloading and running executables) without explicit user consent, which is a common malware distribution pattern and therefore suspicious.

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md instructs the agent to use the BrowserAct CLI to “access real browser pages” and “capture browser-visible page content” at runtime, which can include outsider-authored free text from arbitrary public web pages the user targets.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to install and invoke the BrowserAct CLI/Skill at runtime using the URL(s) https://www.browseract.com/?co-from=ecommerce and https://www.browseract.com/?co-from=ecommerce&redirect=https://github.com/browser-act/skills/tree/main, which are required runtime dependencies used to fetch and execute remote tooling that controls the agent's browser workflow.