Context Leakage
High
- Category
- Data Exfiltration
- Confidence
- 75% confidence
- Finding
Code or instructions that leak agent conversation context to external services, potentially exposing sensitive user interactions.
- Content
md 1. **Check for urgent red flags first.** If the user describes sudden severe headache, neurological symptoms, fever, head injury, pregnancy-related concerns, new or worsening pattern, or other alarming signs, advise urgent local medical care before routine logging. 2. **Define the tracking goal.** Clarify whether the user wants a daily template, an episode-by-episode log, a weekly review, or an appointment summary. 3. **Record each episode.** Capture date, start time, end time, severity, pain location, symptom details, disruption level, and what the user was doing when it began. 4. **Capture context factors.** Add fields for sleep, meals, hydration, caffeine, alcohol, stress, screens, weather, cycle, exercise, travel, illness, and unusual routine changes. 5. **Track interventions and response.** Record medication or other steps used, dose only if the user provides it, timing, relief level, side effects, and whether symptoms returned. 6. **Review repeated patterns cautiously.** Identify repeated associations and unknowns without claiming causation or telling the user what caused the migraine. 7. **Create an appointment summary.** Condense frequency, duration, severity range, common symptoms, possible patterns, treatments tried, response, red flags, and top questions.