hippo-memory@1.38.0
Biologically-inspired memory system for AI agents. Decay by default, strength through use.
AI Security Review
scanned 14d ago · by lpm-firewall-aiSource inspection found an install-time repair limited to an already opted-in Codex wrapper and agent hook setup reached through the explicit CLI init command. The flagged project, prediction, capture, embedding, and hook files provided no concrete attack behavior in the inspected excerpts.
Decision evidence
public snapshot- package.json declares a postinstall wrapper that loads dist/postinstall.js; the postinstall code repairs a Codex wrapper only for users who previously installed it.
- Agent hook installation is called by the explicit hippo init command and can be disabled with --no-hooks.
- The init hook detector checks for agent-related project files, and the implementation patches an instruction file only when it already exists.
- The other flagged files contain project identity, capture, prediction, provider, and hook declarations or paths; the inspected excerpts show no concrete malicious behavior.
- No inspected source shows credential theft, hidden network exfiltration, destructive behavior, or an automatic first-time agent configuration change during package installation.
Source & flagged code
11 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 a known benign dynamic code generation pattern.
dist/rerankers/cross-encoder.jsView on unpkg · L9Findings
Affected versions and remediation
This report applies to hippo-memory@1.38.0.
See all versions of hippo-memory for other recorded verdicts.
- Review the evidence and your use of hippo-memory@1.38.0 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.