google-flow-prompting
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of instructional Markdown content and documentation. No scripts, binaries, or executable commands are included in the file.
- [DATA_EXFILTRATION]: No sensitive data access, credential harvesting, or network-based exfiltration patterns were detected. There are no hardcoded secrets or references to sensitive environment variables.
- [EXTERNAL_DOWNLOADS]: The skill does not perform any external downloads or package installations. References to official documentation and external services are informational only.
- [PROMPT_INJECTION]: The instructions do not contain attempts to bypass model safety filters, override system instructions, or extract internal system prompts. The guidance focuses on legitimate model steering through structured prompting.
- [OBFUSCATION]: No obfuscated strings, hidden Unicode characters, or multi-layer encoding techniques were found in the analyzed content.
Audit Metadata