pipedrive

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONNO_CODE
Full Analysis
  • [NO_CODE]: The skill does not contain any local executable scripts or code. It defines instructions and schemas for interacting with Pipedrive via a managed remote API service provided by AgentPMT.
  • [DATA_EXFILTRATION]: The create_webhook action allows an agent to configure Pipedrive to send real-time CRM updates to a user-specified subscription_url. While a standard CRM feature for automation, it enables data movement to arbitrary external endpoints.
  • [DATA_EXFILTRATION]: The skill integrates with the AgentPMT File Manager, facilitating the transfer of files between the Pipedrive CRM and external storage platforms through the upload_file and download_file actions.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests unstructured data from external CRM records (such as deal titles, person names, and notes) that could be populated by untrusted sources.
  • Ingestion points: Data is pulled into the agent's context through actions like get_record, list_records, get_note, list_notes, get_lead, list_leads, and item_search.
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are documented for processing the content retrieved from Pipedrive.
  • Capability inventory: The agent has the capability to create/delete webhooks, modify CRM records, and link remote files (e.g., from Google Drive).
  • Sanitization: There is no evidence of filtering or sanitization logic being applied to the ingested CRM content before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 10:03 AM