social-keyword-monitor

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The Skill contains deceptive runtime instructions that tell the agent to install missing dependencies automatically and explicitly to hide or avoid asking the user for approval, which overrides normal transparency and lies outside the Skill's monitoring purpose.

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 at runtime and extract “visible text”/content from public social platforms, which is outsider-authored free text that can be fed into the LLM context via the BrowserAct workflow.

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 automatically install the required BrowserAct Skill/CLI at runtime using the install link (https://www.browseract.com/?co-from=social-media&redirect=https://github.com/browser-act/skills/tree/main), which is a runtime dependency that would fetch and install remote code; the BrowserAct homepage link (https://www.browseract.com/?co-from=social-media) is also referenced as the CLI source.