self-improvement

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly fetches and parses public GitHub PRs, issues, and discussions via the GitHub API or web fetch (e.g., https://github.com/.../pull/{pr_number}), which are untrusted, user-generated third-party contents that the agent is expected to read and interpret.