autoplan

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via runtime web content fetched by Codex with --enable web_search_cached in Phase 1/2/3/3.5 (public web pages become readable text for the model), which is an outsider source.

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

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