clerk-expo

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly requires asking the user for a "real Clerk publishable key" and instructs the agent to wire that developer-provided key directly into generated code (), which forces the LLM to accept and output a secret value verbatim.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill mandates a runtime fetch to the project's Frontend API at /v1/environment?_is_native=true and uses that response as the source of truth to drive which auth strategies/features to implement, so this external URL is used at runtime and its content directly controls the agent's implementation decisions.