Caveat verdict
openclaw-memory-core
Shared library for OpenClaw memory plugins providing local redaction, JSONL storage, and offline hash-based embeddings; all operations are local with explicit design to prevent secret leakage.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Permission integrity
package_install
Findings (12)
Accesses sensitive system files
README.md · code · /etc/passwd
Prompt injection — tries to override agent instructions
tests/injection.test.ts · prose · downgraded · IGNORE ALL PREVIOUS INSTRUCTIONS
<script> tag in markdown — potential code injection
tests/injection.test.ts · prose · downgraded · <script>
Possible hardcoded credential
tests/injection.test.ts · prose · downgraded · secret = "sk-abcdefghijklmnopqrstuvwxyz1234567890
Recursive delete from root or home — destructive command
tests/injection.test.ts · prose · downgraded · rm -rf /
Accesses sensitive environment variables
README.md · code · process.env.OPENAI_API_KEY
Instructs covert action — may act without user awareness
src/utils.ts · prose · downgraded · silently
Possible prompt injection — attempts to redefine agent identity
tests/injection.test.ts · prose · downgraded · You are now
Dynamic import() — loads module at runtime
tests/injection.test.ts · prose · downgraded · import("
Makes HTTP request to external URL
README.md · code · fetch("https://
Popular HTTP library — network access
src/utils.ts · prose · downgraded · Got
Base64 encoding/decoding
tests/injection.test.ts · prose · downgraded · base64-encode
Permissions & capabilities
No declared permissions — minimal attack surface.
package_install Is this flag fair?
Thanks — recorded.