@harness-memory/openclaw
OpenClaw native memory plugin backed by harness-memory (SQLite + FTS5, atom/tree/raw recall).
This report shows the evidence for @harness-memory/openclaw@0.2.1. 1 of 3 known versions is blocked or flagged. The newest version, 0.2.2, is allowed. See all versions.
AI Security Review
scanned 3mo ago · by lpm-firewall-aiThis is a concrete, unconsented install-time mutation of a foreign AI-agent control surface, even though the behavior is documented and conditional. No network exfiltration or concealed payload was found.
Decision evidence
public snapshot- package.json runs scripts/postinstall.mjs on npm install.
- scripts/postinstall.mjs probes executables then invokes `harness-memory openclaw setup`.
- The postinstall comments and output state setup writes OpenClaw's openclaw.json.
- dist/setup.js sets plugins.slots.memory to harnessmemory, replacing an existing memory slot.
- Postinstall is conditional on an installed bridge and CLI, and can be disabled with HARNESSMEMORY_SKIP_POSTINSTALL=1.
- The configuration behavior is openly documented in README.md; no hidden network client or payload loader was found.
- Runtime bridge communication is local child-process stdin/stdout JSON-RPC.
Source & flagged code
2 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 unpkgFindings
Affected versions and remediation
This report applies to @harness-memory/openclaw@0.2.1.
The versions list below shows the verdict for each recorded version.
- Avoid installing @harness-memory/openclaw@0.2.1. 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.