revyl-cli-auth-bypass

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of deep link handlers (e.g., myapp://revyl-auth) that process external parameters including tokens, roles, and redirect paths.
  • Ingestion points: Deep link URL handling is defined in SKILL.md and detailed in references for iOS, Android, Flutter, and Expo.
  • Boundary markers: The skill requires implementations to be gated by a specific launch variable (REVYL_AUTH_BYPASS_ENABLED) and restricted to non-production builds.
  • Capability inventory: The agent is granted the ability to configure project files (.revyl/config.yaml), manage environment variables via CLI, and interact with test devices (navigation, screenshots).
  • Sanitization: Instructions explicitly mandate allowlisting for roles and redirect routes and require token validation before the application state is modified.
  • [COMMAND_EXECUTION]: The skill relies on the revyl CLI and standard shell utilities to perform its tasks.
  • The agent is instructed to execute commands such as revyl init, revyl dev, and revyl global launch-var to manage the testing environment.
  • File system discovery is performed using find, ls, and pwd to detect the mobile application stack (e.g., searching for Podfile, build.gradle, or pubspec.yaml).
  • The before_script feature in the Revyl configuration allows for the execution of repository-local scripts (e.g., ./scripts/prepare-test-session.sh) to automate token generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 09:39 PM