@expo/agent-cli@1.0.22
Agent-native CLI on top of the Expo CLI family
AI Security Review
scanned 6d ago · by lpm-firewall-aiThe CLI sends runtime command telemetry by default, including detected agent and sandbox context, without documenting that automatic collection in its README. This is undisclosed runtime telemetry, so the package is suspicious rather than malicious.
Decision evidence
public snapshot- Each CLI command starts a detached telemetry worker unless telemetry is disabled or Expo offline mode is active.
- The worker sends command telemetry to https://cdp.expo.dev/v1/batch and includes runtime, CI, agent, and sandbox context.
- The README documents metadata collection and opt-outs for the feedback command, but does not disclose the automatic command telemetry.
- The automatic telemetry payload contains the command name and package version; the inspected code does not show command arguments or output being sent.
- Agent instruction-file updates are implemented as a CLI setup operation, rather than an install hook.
Source & flagged code
11 flaggedPackage source references child process execution.
build/cli/8796.index.jsView on unpkg · L97Source contains an obfuscated payload loader that reconstructs and executes hidden code.
build/cli/4108.index.jsView on unpkg · L88192Findings
Affected versions and remediation
This report applies to @expo/agent-cli@1.0.22.
See all versions of @expo/agent-cli for other recorded verdicts.
- Review the evidence and your use of @expo/agent-cli@1.0.22 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.