Caveat verdict
Memoria for OpenClaw
openclaw-memoria
Local-first SQLite memory system with optional remote LLM fallback keys declared transparently; security section explicitly scopes file access to workspace only.
⚠ 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 (11)
Pipe-to-shell pattern (curl | sh) — supply chain attack vector
INSTALL.md · code · curl -fsSL https://raw.githubusercontent.com/Primo-Studio/openclaw-memoria/main/
Pipe to bash — executes piped content as shell commands
INSTALL.md · code · | bash
Pipe to sh — executes piped content as shell commands
core/procedural.ts · prose · downgraded · |sh
Instructs covert action — may act without user awareness
CHANGELOG.md · prose · downgraded · silently
Dynamic import() — loads module at runtime
continuous.ts · prose · downgraded · import("
Uses exec() — may execute shell commands
core/db.ts · prose · downgraded · exec(
Accesses sensitive environment variables
core/README.md · code · process.env.OPENAI_API_KEY
Python os.environ.get — reads environment variable
benchmarks/bench-v34.py · prose · downgraded · os.environ.get(
Blob URL — may embed executable content
core/embeddings.ts · prose · downgraded · blob:
References agent memory files
core/md-regen.ts · prose · downgraded · MEMORY.md
Popular HTTP library — network access
core/procedural.ts · prose · downgraded · got
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A/B). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
No declared permissions — minimal attack surface.
package_install Is this flag fair?
Thanks — recorded.