x-tweet-by-conversation

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

• Third-party content exposure detected (high risk: 0.85). Yes—runtime network capture of X’s GraphQL TweetDetail responses (containing tweet text and author bios/descriptions authored by other users) is saved as readable JSON/text files and then parsed by scripts/parse-tweets.py, whose extracted fields (e.g., text, author.description) are emitted into the agent’s LLM context as structured output.