pnpm-upgrade

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly instructs fetching untrusted public content—e.g., "curl -fsSL https://api.github.com/repos/pnpm/action-setup/releases/latest" and using git ls-remote against the pnpm/action-setup repo—then uses those release/tag values to choose commits and update CI workflow pins, so third-party release data directly influences actions the agent takes.