Ops Automation Builder

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a structured framework for automation (triggers, conditions, actions) that operates on internal data tables.
  • [DATA_EXPOSURE]: While the skill interacts with external services like HubSpot and Slack, these are intended functional integrations. The skill does not hardcode credentials or exfiltrate data to unauthorized domains.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes data from table cells which may originate from untrusted sources.
  • Ingestion points: Data enters the system through table_id cell values, row creation events, and enrichment results.
  • Boundary markers: The skill uses structured JSON for logic definitions (condition, action_config), which helps isolate data from instructions, though explicit boundary markers for cell content are not specified.
  • Capability inventory: The skill can perform write operations (update_cell, add_tag), execute external tools (run_enrichment), and perform network operations (push_to_hubspot, notify).
  • Sanitization: There is no explicit mention of sanitizing or escaping external content before processing or interpolation into notification messages.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:37 AM