granola-sdk-patterns
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides integration patterns and documentation for connecting Granola with well-known, trusted services including Zapier, Notion, Slack, and HubSpot.
- [PROMPT_INJECTION]: The skill describes automation workflows that ingest and process untrusted meeting data, creating a surface for indirect prompt injection.
- Ingestion points: Workflow templates in
SKILL.mdingest meeting content through placeholders such as{{note_content}},{{summary}}, and{{action_items}}. - Boundary markers: There are no explicit boundary markers or delimiters defined in the templates to prevent the agent from obeying instructions potentially hidden in the ingested meeting notes.
- Capability inventory: The skill specifies access to
Bash(curl:*),Read,Write, andEdittools in its YAML frontmatter. - Sanitization: No sanitization or filtering of the external meeting content is described before it is interpolated into downstream actions, such as the OpenAI email generation step.
Audit Metadata