gh-issue-to-pr

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly loads and parses GitHub issue text with the command gh issue view ... in "Step 1: Verify and Load Context" of SKILL.md, meaning it ingests untrusted, user-generated content from public GitHub issues that can influence implementation decisions and tool use.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs git/gh at runtime to fetch remote repositories (e.g., "gh repo clone <owner/repo>" and "git remote add upstream https://github.com/<owner/repo>.git"), and those fetched repos are then built/tested/ran as part of the workflow, so remote code can be executed and thus can directly control agent behavior.