@tekmidian/pai@0.60.0
PAI Knowledge OS — Personal AI Infrastructure with federated memory and project management
AI Security Review
scanned 11d ago · by lpm-firewall-aiSource shows guarded, explicit setup for package-owned Claude hooks and settings, which is a real agent-configuration capability but does not establish unconsented control-surface mutation. The manifest has no install-time hook, and the inspected credential and notification excerpts do not prove exfiltration; classify as suspicious under the guarded agent integration policy.
Decision evidence
public snapshot- The manifest has a prepublish build hook and no install-time lifecycle hook.
- The explicit setup flow offers to install package-owned lifecycle hooks and patch Claude settings with hook registrations, permissions, and flags.
- The setup flow asks before updating the user’s Claude instructions; this is user-invoked, guarded configuration behavior.
- Todoist setup asks before using an environment token, and the inspected hook excerpt only reads an optional notification topic.
- The available excerpts do not establish credential theft or unconsented agent-control mutation.
- Setup can add package hooks and permissions to Claude settings, creating a guarded agent configuration capability.
Source & flagged code
13 flagged · loading sourcePackage source references child process execution.
dist/daemon-DArRKBlk.mjsView on unpkg · L16Findings
Affected versions and remediation
This report applies to @tekmidian/pai@0.60.0.
See all versions of @tekmidian/pai for other recorded verdicts.
- Review the evidence and your use of @tekmidian/pai@0.60.0 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.