memoryrouter-codex@1.8.1
Automatic MemoryRouter memory for OpenAI Codex: one vault shared by every session, repo and tool, including Claude Code
AI Security Review
scanned 1d ago · by lpm-firewall-aiThe package has no install lifecycle hook, and its Codex configuration changes occur through the explicit `init` command. Its runtime sends conversation content to MemoryRouter for the advertised memory feature; the reviewed credential and process flows serve that purpose.
Decision evidence
public snapshot- The manifest declares no install, preinstall, or postinstall lifecycle script.
- Writes to Codex configuration and agent instructions run through the explicit `init` command.
- The package describes its hooks as saving the user's message and final response to MemoryRouter; it excludes tool calls and output.
- Runtime API requests send the configured key to the MemoryRouter API base.
- Automatic consolidation reuses the package's reflection flow and invokes Codex with retrieved memories.
- Hooks automatically capture conversation content after setup, but this is disclosed and central to the package's stated memory feature.
- The package uses child processes for Git identity lookup and Codex-based consolidation; the reviewed paths tie these actions to project labeling and the memory feature.
Source & flagged code
6 flaggedPackage source references child process execution.
runtime/project.cjsView on unpkg · L29A single source file combines environment access, network access, and code or shell execution; review context before blocking.
lib/installer.cjsView on unpkg · L174Findings
Affected versions and remediation
This report applies to memoryrouter-codex@1.8.1.
See all versions of memoryrouter-codex for other recorded verdicts.
- Review the evidence and your use of memoryrouter-codex@1.8.1 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.