Caveat verdict
agent-brain
Local-first SQLite memory system with pluggable backends; all storage is local with optional SuperMemory mirroring that is user-configured, and no concrete evidence of unauthorized data exfiltration is present in the content.
⚠ 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.
Findings (9)
Instructs covert action — may act without user awareness
modules/archive/SKILL.md · code · silently
Possible hardcoded credential
scripts/memory.sh · prose · downgraded · API_KEY="$parsed_key
Uses eval() — can execute arbitrary code
scripts/test_memory.sh · prose · downgraded · eval(
Pipe to python — executes piped content as Python code
scripts/test_memory.sh · prose · downgraded · | python3
Popular HTTP library — network access
modules/signal/SKILL.md · code · Got
Gopher protocol — SSRF attack vector
modules/ingest/SKILL.md · prose · downgraded · gopher://
HTTP request to bare IP address — common in malicious payloads
scripts/test_memory.sh · prose · downgraded · http://127.0.0.1
Python urllib.request — network access
scripts/brain.py · prose · downgraded · urllib.request
Python os.environ.get — reads environment variable
scripts/brain.py · prose · downgraded · os.environ.get(
Permissions & capabilities
No declared permissions — minimal attack surface.
Is this flag fair?
Thanks — recorded.