domma-js@0.51.7
Dynamic Object Manipulation & Modeling API - A complete front-end toolkit.
AI Security Review
scanned 11d ago · by lpm-firewall-aiThe install-time agent configuration setup is guarded by an interactive prompt, avoids overwriting existing destinations, and is also available through a user-invoked CLI command. This is a guarded first-party agent setup capability, warranting a warning rather than a block; the scanner's malware label is not supported by the inspected source.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review
Trigger
Installing the package under node_modules outside a CI environment, followed by an affirmative response or accepting the default yes.
Impact
The project may gain package-provided Claude Code guidance and settings. No exfiltration, remote execution, or overwrite behavior was established from the inspected source.
Mechanism
The hook locates the project root and copies the package's CLAUDE.md and .claude directory only when those destinations are absent; the CLI also offers an explicit setup-ai command.
AI rationale
The install-time agent configuration setup is guarded by an interactive prompt, avoids overwriting existing destinations, and is also available through a user-invoked CLI command. This is a guarded first-party agent setup capability, warranting a warning rather than a block; the scanner's malware label is not supported by the inspected source.
Evidence
package.jsonbin/postinstall.jsbin/domma-cli.jsCLAUDE.md.claude/
Decision evidence
public snapshotAI called this Suspicious at 91.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The package runs bin/postinstall.js automatically after installation.
- The hook gates itself on installation under node_modules and skips CI environments.
- The hook asks whether to copy AI assistance files; answering no cancels the operation.
- After consent, the hook copies package-provided Claude Code files only when corresponding project paths do not already exist.
- The CLI exposes setup-ai as a separate command and prompts before copying its AI assistance files.
- The flagged view template only logs route and mount events; its request example is commented out.
Evidence against
- The postinstall hook reaches the consumer project root and can create CLAUDE.md and .claude configuration during dependency installation after an affirmative or default-yes response.
Behavioral surface
ChildProcessEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsMinified
Source & flagged code
7 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node bin/postinstall.js
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node bin/postinstall.js
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgpublic/dist/domma.esm.jsView file
9patternName = generic_password
severity = medium
line = 9
Medium
bin/postinstall.js
Findings
1 Critical3 High6 Medium4 Low
CriticalAi Agent Control Hijackbin/postinstall.js
HighInstall Time Lifecycle Scriptspackage.json
HighSemantic Analysis Limitedtemplates/view-template/view.js
HighKnown Malware Source Similaritybin/domma-cli.js
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumSecret Patternpublic/dist/domma.esm.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
MediumSecret Patternpublic/dist/domma.min.js
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to domma-js@0.51.7.
See all versions of domma-js for other recorded verdicts.
- Review the evidence and your use of domma-js@0.51.7 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.