memoryrouter-codex@1.2.0
Automatic, project-scoped MemoryRouter memory for OpenAI Codex
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThe package is not malicious by static inspection, but its explicit setup installs persistent Codex hooks that transmit conversation content to a remote service. This warrants a warning for agent extension lifecycle risk rather than a publication block.
Decision evidence
public snapshot- An explicit init command writes MCP configuration, hook configuration, and agent guidance into Codex-controlled locations.
- Installed hooks run on several Codex lifecycle events and capture completed prompts and assistant responses.
- Captured content is sent with a bearer token to the MemoryRouter API.
- Project API-key configuration can invoke the package again through npx at runtime.
- The manifest has no preinstall, install, or postinstall lifecycle script.
- Configuration changes require the user to run the init command.
- The hook does nothing when no API key is configured.
- No stealthy code loading, credential harvesting beyond the configured API key, or destructive behavior was found.
Source & flagged code
5 flaggedPackage source references child process execution.
runtime/project.cjsView on unpkg · L4A single source file combines environment access, network access, and code or shell execution; review context before blocking.
lib/installer.cjsView on unpkg · L124Findings
Affected versions and remediation
This report applies to memoryrouter-codex@1.2.0.
See all versions of memoryrouter-codex for other recorded verdicts.
- Review the evidence and your use of memoryrouter-codex@1.2.0 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.