salary-research

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md Research procedure explicitly instructs fetching and ingesting public third‑party sources (e.g., job postings and salary platforms like Levels.fyi, Glassdoor, Payscale and anonymous community sources such as Blind/Fishbowl) which the agent will read and use to drive benchmarking outputs, so untrusted/user‑generated content can influence its decisions.