evals@2.7.0
Arize AX onboarding — instrument your app with tracing via your coding agent
AI Security Review
scanned 3mo ago · by lpm-firewall-aiThis is a consent-gated, first-party agent-extension setup with a concrete telemetry collection surface, not covert install-time malware. Warn so users receive visibility before enabling it.
Decision evidence
public snapshot- Vendored tracing wheel installs agent settings/hooks and captures session data.
- Agent-tracing prompt explicitly targets every session and tool outputs.
- Harness installation is invoked only through the user-run onboarding flow.
- postinstall only prints an installation message.
- CLI stages a bundled prompt, then launches a user-selected local agent.
- Prompt requires a separate explicit yes before machine-wide tracing.
- No install-time payload execution, secret harvesting, or exfiltration was found.
Source & flagged code
7 flaggedPackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to evals@2.7.0.
See all versions of evals for other recorded verdicts.
- Review the evidence and your use of evals@2.7.0 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.