Caveat verdict
openclaw-bastion
Prompt injection defense scanner analyzing local workspace files for malicious patterns; all operations are read-only local analysis with no external network calls, matching the stated security scanning purpose.
⚠ 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)
Recursive delete from root or home — destructive command
README.md · prose · downgraded · rm -rf /
Raw model control tokens — prompt injection via token manipulation
scripts/bastion.py · prose · downgraded · <|im_start|>
<script> tag in markdown — potential code injection
scripts/bastion.py · prose · downgraded · <script>
Bash /dev/tcp — raw TCP connection via shell
scripts/bastion.py · prose · downgraded · /dev/tcp/
References agent memory files
scripts/bastion.py · prose · downgraded · MEMORY.md
References agent configuration files
scripts/bastion.py · prose · downgraded · CLAUDE.md
Sets world-executable permissions
scripts/bastion.py · prose · downgraded · chmod 777
Python os.environ.get — reads environment variable
scripts/bastion.py · prose · downgraded · os.environ.get(
Python shutil file operation — copies/moves/deletes files
scripts/bastion.py · prose · downgraded · shutil.move(
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
Requires 1 system binary.
Is this flag fair?
Thanks — recorded.