create-agent

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands for project initialization, dependency management, and building. Evidence: Phase 3 of SKILL.md includes instructions to run npx eve@latest init, pnpm install, and pnpm build.
  • [EXTERNAL_DOWNLOADS]: The skill downloads the framework and external packages from public registries. Evidence: SKILL.md specifies the use of npx and pnpm to fetch the 'eve' package and other project dependencies.
  • [PROMPT_INJECTION]: The skill implements an indirect prompt injection surface by interpolating user-provided answers from a design interview directly into the generated agent's instructions. Ingestion points: User responses during the Phase 1 Design interview documented in SKILL.md. Boundary markers: None are specified to delimit user-provided text within the generated agent/instructions.md file. Capability inventory: The skill possesses the ability to perform file writes and execute shell commands. Sanitization: There is no evidence of input validation or escaping for the user responses before they are written to the instructions file.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 07:15 AM
Security Audit — agent-trust-hub — create-agent