granola-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of instructional documentation for a SaaS workflow. It does not contain any executable scripts, binary files, or automated commands that could pose a security risk.
- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow where the agent processes meeting transcripts. While transcripts are external data sources that could theoretically contain hidden instructions, the skill implements a manual review step (Step 2) and relies on the platform's native LLM processing, which is standard for productivity tools.
- [DATA_EXFILTRATION]: The sharing functionality mentioned (Slack, Notion, CRM) refers to native SaaS integrations. There are no patterns suggesting unauthorized exfiltration, hardcoded credentials, or network requests to untrusted third-party domains.
- [EXTERNAL_DOWNLOADS]: All external resource links point to official Granola documentation and blog posts (granola.ai), which are trusted sources for the platform's own skill.
Audit Metadata