Caveat verdict
ironclaw-security-guard
Security guardrails adding shell-risk filtering, prompt-injection detection, and secret redaction to OpenClaw; execution sink is a local defensive security scan tool.
⚠ 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 (8)
Recursive delete from root or home — destructive command
README.md · code · rm -rf /
Dynamic import() — loads module at runtime
README.md · code · import('
Pipe-to-shell pattern (curl | sh) — supply chain attack vector
README.zh-CN.md · prose · downgraded · curl ... | sh
Prompt injection — tries to override agent instructions
README.zh-CN.md · prose · downgraded · ignore previous instructions
Pipe to sh — executes piped content as shell commands
README.zh-CN.md · prose · downgraded · | sh
Accesses sensitive system files
src/config.ts · prose · downgraded · /etc/shadow
Pipe to bash — executes piped content as shell commands
src/config.ts · prose · downgraded · |bash
Accesses system credential store
src/config.ts · prose · downgraded · Keychain
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A). 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.
Is this flag fair?
Thanks — recorded.