@hashbrownai/openai@0.8.3
OpenAI provider for Hashbrown AI
AI Security Review
scanned 9d ago · by lpm-firewall-aiSource inspection confirms default postinstall disclosure of installation details and potentially Git identity to the publisher’s collector, with opt-outs but no affirmative consent gate. The package-aligned recipient and lack of other confirmed attack behavior support a warning rather than a publish block.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review
Trigger
Installing the package with lifecycle scripts enabled.
Impact
The collector may receive an installation identifier, environment details, and Git display name or email without affirmative installer consent.
Mechanism
The postinstall hook gathers installation and checkout details, builds an event, and posts it to the default Hashbrown collector unless an opt-out is set.
AI rationale
Source inspection confirms default postinstall disclosure of installation details and potentially Git identity to the publisher’s collector, with opt-outs but no affirmative consent gate. The package-aligned recipient and lack of other confirmed attack behavior support a warning rather than a publish block.
Evidence
package.jsoninstall/postinstall.cjsinstall/collector.cjsinstall/policy.cjsinstall/identity.cjsinstall/git-context.cjs
Network endpoints1
hashbrown.dev/_/growth/collect/v1/install
Decision evidence
public snapshotAI called this Suspicious at 88.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The manifest runs an install hook that starts the collector during postinstall.
- The collector sends an installation event containing a persistent installation identifier and package, operating system, architecture, and Node details.
- Outside CI, the event can include configured Git display name and email; no affirmative consent gate precedes collection or sending.
- The default recipient is the Hashbrown-branded collector, and endpoint overrides are restricted to loopback.
- Telemetry can be disabled with opt-out flags, but collection is enabled by default.
Evidence against
- The recipient matches the package publisher’s Hashbrown identity, and the source does not show an unrelated destination or a separate malicious payload.
- The collector honors telemetry opt-outs and omits Git identity in CI.
Behavioral surface
CryptoEnvironmentVarsFilesystemNetwork
UrlStrings
Source & flagged code
7 flaggedpackage.jsonView file
•scripts.postinstall = node install/postinstall.cjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node install/postinstall.cjs
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkg37"postinstall": "node install/postinstall.cjs"
}
}
Medium
Suspicious Lifecycle Evidence
The manifest runs an install hook that starts the collector during postinstall.
package.jsonView on unpkg · L37install/git-context.cjsView file
Findings
3 High6 Medium3 Low
HighInstall Time Lifecycle Scriptspackage.json
HighKnown Malware Source Similarityinstall/git-context.cjs
HighKnown Malware Source Similarityinstall/policy.cjs
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumNetwork
MediumEnvironment Vars
MediumSuspicious Lifecycle Evidencepackage.json
MediumSuspicious Lifecycle Evidenceinstall/postinstall.cjs
MediumAi Review Evidenceinstall/collector.cjs
LowScripts Present
LowFilesystem
LowUrl Strings
Affected versions and remediation
This report applies to @hashbrownai/openai@0.8.3.
See all versions of @hashbrownai/openai for other recorded verdicts.
- Review the evidence and your use of @hashbrownai/openai@0.8.3 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.