expo-skill-feedback

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and run the submit-expo-feedback package from npm using the command npx --yes submit-expo-feedback@latest.
  • [REMOTE_CODE_EXECUTION]: Execution of the remote submit-expo-feedback package via npx is used to submit user-approved feedback and failure reports (eval candidates) to Expo's backend.
  • [DATA_EXFILTRATION]: The skill includes a telemetry system that sends anonymous usage data, such as OS name, architecture, a hashed installation ID, and the name of the invoked skill, to PostHog (us.i.posthog.com). This telemetry is explicitly off by default and requires user opt-in via a provided script or environment variables.
  • [COMMAND_EXECUTION]: The telemetry script scripts/skill-event.cjs uses child_process.spawn to launch a detached background process to handle network requests without blocking the agent's interaction flow.
  • [CREDENTIALS_UNSAFE]: The telemetry configuration in scripts/telemetry_common.cjs includes a hardcoded PostHog project API key. This is documented as a public, write-only ingestion key intended for client-side data submission.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 01:42 PM
Security Audit — agent-trust-hub — expo-skill-feedback