github-daily-brief

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

• Third-party content exposure detected (high risk: 0.90). The skill instructs the agent to use the global agent-browser to fetch public GitHub pages (e.g., https://github.com/trending?since=daily and search results like the created:YYYY-MM-DD URL, plus README/topics) and to read/interpret those user-generated descriptions to produce summaries and recommendations, which means untrusted third-party content directly influences the agent's outputs and decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill uses agent-browser at runtime to fetch GitHub pages (e.g. https://github.com/trending?since=daily and https://github.com/search?q=created%3A2026-02-11&sort=stars&order=desc&type=repositories), and that fetched repository/README content is injected into the model context to drive the LLM’s summaries, so external content directly controls the agent’s output and is required for the skill.