thesis-docx

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md:24-25 instructs runtime loading of “Wiki knowledge” via maestro search ... --json, which is public/outsider-authored web content and can be ingested as free text into the agent’s LLM context.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). The skill explicitly runs PowerShell with "-ExecutionPolicy Bypass" (SKILL.md:31-33), which instructs bypassing a Windows security mechanism, so it is flagged as a potential compromise of machine state.