@adithyasak/t3@0.1.2-adi.1
Node.js WebSocket server for the T3 Code app.
AI Security Review
scanned 12d ago · by lpm-firewall-aiThe package manifest exposes a CLI and contains no lifecycle scripts. The scanner’s shell, environment, and dynamic-code matches resolve to bundled agent functionality and an empty-function capability probe; inspected source does not establish malicious behavior.
Decision evidence
public snapshot- package.json publishes a CLI through dist/bin.mjs and declares no install or import lifecycle hook.
- The Claude history worker handles requested message retrieval or session forking; the CLI passes it the requested operation, session ID, and options.
- The new Function match is a capability check that tries an empty function body and catches failure; it does not construct or run a payload.
- The worker’s environment and shell code creates a Bash session for CLI functionality, which explains the process and environment matches.
- The flagged Effect files contain standard runtime, collection, encoding, schema, HTTP-data, and networking helpers; the inspected dependency manifest identifies them as Effect 4.0.5.
- The other inspected dependency paths identify the Claude Agent SDK and Cursor SDK; the Claude SDK export includes session history and fork APIs used by the worker.
Source & flagged code
12 flaggedSource appears to send environment or credential material to an external endpoint.
dist/claudeHistoryWorker-Dh6fwOGe.mjsView on unpkg · L11Source matches reverse-shell style process and socket wiring.
dist/claudeHistoryWorker-Dh6fwOGe.mjsView on unpkg · L11Findings
Affected versions and remediation
This report applies to @adithyasak/t3@0.1.2-adi.1.
See all versions of @adithyasak/t3 for other recorded verdicts.
- Review the evidence and your use of @adithyasak/t3@0.1.2-adi.1 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.