ClawAudit verdict
deep-research
ars-deep-research
Academic research orchestrator that delegates to sub-agents via delegate_task for literature review and synthesis; the pipeline is transparent and matches the stated purpose of deep research with no exfiltration patterns.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. ClawAudit flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Findings (3)
Possible hardcoded credential
shared/cross_model_verification.md · code · API_KEY="sk-your-key-here
Instructs covert action — may act without user awareness
agents/report_compiler_agent.md · prose · downgraded · silently
Accesses shell history/config
shared/cross_model_verification.md · prose · downgraded · ~/.zshrc
Permissions & capabilities
No declared permissions — minimal attack surface.
Is this flag fair?
Thanks — recorded.