moai-workflow-docs-claim-check
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEPROMPT_INJECTIONCREDENTIALS_UNSAFEDATA_EXFILTRATION
Full Analysis
- [SAFE]: The skill is designed for read-only document analysis and includes explicit instructions to avoid command execution and protect sensitive data found in processed documents.\n- [PROMPT_INJECTION]: The skill uses clear instructions and "hard boundaries" to restrict agent behavior, specifically forbidding the execution of shell commands, network requests, or file writes.\n- [CREDENTIALS_UNSAFE]: The skill includes a "Preflight" step specifically designed to scan for and flag secrets (credentials, tokens, private keys) in the evidence documents to prevent them from being reproduced in the output.\n- [DATA_EXFILTRATION]: No network operations or external data transfer patterns were identified. The skill is restricted to Read, Grep, and Glob tools for local file analysis.
Audit Metadata