code-e2e-tests

Pass

Audited by Gen Agent Trust Hub on Jul 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands such as npm run cypress:run and npm run cypress:open. These are standard commands used to run and debug end-to-end tests within a Node.js environment.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing external application data and task requirements to generate executable test code.
  • Ingestion points: Reads frontend source code (frontend/), existing command files, and task contract files (e.g., task-*.md as seen in evals).
  • Boundary markers: None explicitly defined to delimit external data from generated logic.
  • Capability inventory: The skill can write TypeScript files (.ts, .cy.ts) and execute shell commands (npm run).
  • Sanitization: No specific sanitization or validation of the ingested external content is described before interpolation into test code.
  • [COMMAND_EXECUTION]: The skill involves dynamic code generation where it writes Cypress test specifications and then executes them via the test runner. This is a primary function of the skill and follows standard development patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 24, 2026, 08:18 PM
Security Audit — agent-trust-hub — code-e2e-tests