open-walnut@0.6.6
Personal AI powered by Claude — task management, Claude Code sessions, memory system, and web UI
AI Security Review
scanned 9d ago · by lpm-firewall-aiThe postinstall silently alters a foreign AI agent’s control prompt by removing its file-read reminder. This is a concrete install-time AI agent control hijack.
Decision evidence
public snapshot- The package runs its postinstall script during installation; that script resolves the installed Claude Agent SDK and writes a replacement to its cli.js.
- The shipped diff removes a conditional system reminder from the SDK’s file-read results by replacing it with an empty string. The postinstall applies the diff without an affirmative consent check.
- The flagged runtime excerpts show local config and cache path setup, a CLI config read, or discovery of local AGENTS.md and skills files; they do not show agent settings being changed.
Source & flagged code
26 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 unpkgPackage source references child process execution.
dist/plugin-store/walnut-rhythm/dist/server.mjsView on unpkg · L8Findings
Affected versions and remediation
This report applies to open-walnut@0.6.6.
See all versions of open-walnut for other recorded verdicts.
- Avoid installing open-walnut@0.6.6. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.