marketwatch

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

• Third-party content exposure detected (high risk: 0.85). At runtime the skill fetches public MarketWatch HTML pages (e.g., /investing/stock/{ticker}, /company-profile, /financials/*, /analystestimates, /options, /download-data) and parses their readable text into JSON that is then printed/returned, so outsider-authored free text from MarketWatch is ingested into the agent context.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.70). El skill explícitamente describe técnicas para eludir la protección anti-bot de MarketWatch (uso de headers de navegador completos, manejo de cookies y delays para "bypass" de Datadome), es decir instrucciones para sortear mecanismos de seguridad, por eso lo marco.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)