AI Security Review
scanned 4h ago · by lpm-firewall-aiLPM blocks this version under the AI-agent control-surface policy. The npm postinstall hook mutates the consumer project automatically. It drops package-supplied VS Code MCP agent configuration and instruction files into the host repository.
Decision evidence
public snapshot- `package.json` runs `node ./bin/install-host.js --postinstall`.
- `bin/install-host.js` defaults the target to `INIT_CWD` or current directory.
- Postinstall recursively copies package `.vscode` and `AGENTS.md` into that target.
- `.vscode/mcp.json` registers MCP commands including `token-saver-mcp`, `codegraph`, `gitnexus`, `repomix`, and `graphify`.
- Existing target files are preserved, but absent AI-agent config is created without user confirmation.
- No credential harvesting or network exfiltration was found in reviewed Node install code.
- No remote URL, downloaded payload, `eval`, or shell-string execution appears in `bin/install-host.js`.
- Noninteractive install skips the interactive bootstrap script after scaffolding.
Source & flagged code
5 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgInstall-time source drops package-supplied AI-agent/MCP control files or instructions.
bin/install-host.jsView on unpkg · L1Source fingerprint signature matches a known malicious package signature; route for source-aware review.
bin/install-host.jsView on unpkgPackage ships non-JavaScript build or shell helper files.
scripts/context/build-repomix.ps1View on unpkg