codex-review

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s README indicates it can use web-research/scraping related skills (e.g., tavily-web/exa-search/firecrawl-scraper) and review content from the GitHub repository, so at runtime it likely fetches public web pages or other outsiders’ text and feeds that readable prose into the agent’s LLM context via those tools.