benchmark-due-diligence

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime fan-out uses FACTS (public benchmark facts) as input to collection/verification agents, and those agents perform WebSearch/fetch on that text—so any outsider-authored free text that is retrieved from public web pages can be ingested into the LLM context via the external search results.

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

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