update-turbo

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

• Third-party content exposure detected (high risk: 0.90). The skill fetches and reads upstream skill files from the user's cloned repo remote (e.g., via git -C ~/.turbo/repo fetch and git -C ~/.turbo/repo show /main:skills//SKILL.md) and then has the agent interpret those SKILL.md contents (including launching agents to merge and decide updates), which exposes it to untrusted third-party repository content that can influence tool actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill performs runtime git fetch/pull from the configured repo remote (origin/upstream — e.g. git@github.com:org/repo.git or https://github.com/org/repo.git) and reads the fetched SKILL.md files into agents' prompts and uses pulled files to overwrite/install skills, so remote repository content directly controls agent instructions and supplies executable/installed code.