pipeline-review

Pass

Audited by Gen Agent Trust Hub on Apr 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from multiple external communication channels. * Ingestion points: The skill reads external data via gmail_read_thread, query_granola_meetings, and Attio list-comments (SKILL.md). * Boundary markers: The workflow instructions do not specify the use of delimiters or 'ignore' instructions for content retrieved from these external tools (SKILL.md). * Capability inventory: The agent has access to tools for creating records/tasks in Attio and drafting or sending emails via Gmail, which could be misused if the agent is manipulated by injected instructions (SKILL.md). * Sanitization: There is no explicit requirement for sanitization or validation of the ingested external content before it is used to generate the Risk Report or recommended actions (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 17, 2026, 06:20 AM