setup-project-stack

Pass

Audited by Gen Agent Trust Hub on Sep 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill is designed to install dependencies using standard package managers like npm, pnpm, and yarn.
  • Evidence: SKILL.md, Goal and Write sections. This primary capability is gated by mandatory user consent and security pre-checks.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads existing repository configuration files (e.g., package.json, lockfiles) to determine the project stack, creating a potential vector for indirect injection.
  • Ingestion points: Explore phase in SKILL.md.
  • Boundary markers: The skill uses a rigid multi-wave questioning process and requires user sign-off on the stack profile before any installation occurs.
  • Sanitization: The skill verifies package versions, licenses, and maintenance status before proposing them for installation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 27, 2026, 12:29 PM