prisma-compute

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

• Third-party content exposure detected (high risk: 0.85). The runtime path node prisma-compute/scripts/verify-compute-surface.mjs spawns npx/bunx to execute create-prisma@latest and @prisma/cli@latest --help, and the agent ingests the resulting CLI stdout/stderr (free-form text) from these outsider-authored packages fetched at runtime from the public registry.

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

• Potentially malicious external URL detected (high risk: 0.80). The verify script (prisma-compute/scripts/verify-compute-surface.mjs) spawns package runners (npx or bunx) to invoke remote packages such as "create-prisma@latest" and "@prisma/cli@latest" (e.g., npx --yes @prisma/cli@latest ... / npx --yes create-prisma@latest ...), which fetch and execute code from the npm registry at runtime, so external content is executed and relied on for the skill's behavior.