ory-local-dev

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx -y -p @ory/claude-code to download the vendor's official plugin package from the NPM registry. This is a standard mechanism for distributing development tools.
  • [REMOTE_CODE_EXECUTION]: The execution of npx involves downloading and running executable code from a remote repository (NPM). Since this originates from the author's own verified package, it is consistent with the skill's development purpose.
  • [COMMAND_EXECUTION]: The skill requires running shell commands, including docker info to check the environment and ory-claude subcommands to manage the containerized services.
  • [CREDENTIALS_UNSAFE]: The documentation includes a hardcoded test password (ory-agent-local-dev-password!) for a seeded local identity (agent@ory-local.dev). This is intended for local prototyping and testing of authentication flows.
  • [INDIRECT_PROMPT_INJECTION]: The skill creates an environment that processes data from local authentication services and user-supplied registration inputs.
  • Ingestion points: Data returned from local Ory API endpoints (Gateway, Kratos, Keto) and user input during the registration flow.
  • Boundary markers: None explicitly defined in the provided instructions.
  • Capability inventory: The agent can execute commands via npx and docker to manage the stack.
  • Sanitization: Relies on the inherent validation and sanitization provided by the Ory identity and permission services.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 03:37 AM