prototype

Warn

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions explicitly direct the agent to ensure prototypes are "trivial to run" by providing commands for project task runners and runtimes such as pnpm, python, and bun.
  • [REMOTE_CODE_EXECUTION]: The skill functions by generating executable code (standalone HTML files with logic or framework-specific UI components) and instructing the environment to execute these scripts. This pattern of dynamic script generation followed by execution instructions represents a significant capability that could be misused.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes user-provided logic and UI requirements to generate executable artifacts.
  • Ingestion points: User prompts describing the prototype requirements or design questions in SKILL.md.
  • Boundary markers: The skill lacks explicit boundary markers or instructions to treat user-provided logic descriptions as data rather than instructions.
  • Capability inventory: The skill possesses the ability to write new files to the local file system and execute shell commands to run the generated prototypes.
  • Sanitization: There is no mechanism described for sanitizing or validating user-provided descriptions before they are interpolated into the generated code.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 19, 2026, 05:19 PM
Security Audit — agent-trust-hub — prototype