Pipedrive Automation

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by retrieving data from Pipedrive records and embedding it in communications.
  • Ingestion points: Data enters the agent context from Pipedrive deals, organizations, and persons via the crm-mcp tools.
  • Boundary markers: No delimiters (such as XML tags or explicit escaping) are used in email templates or Slack message configurations to isolate interpolated fields like {{company_insight}} or {{deal.title}}.
  • Capability inventory: The skill possesses the capability to send emails and post to Slack channels, which could be leveraged if malicious instructions are stored in CRM fields.
  • Sanitization: No input validation or content filtering is specified for the CRM data before it is processed.
  • [NO_CODE]: This skill definition is provided entirely in Markdown and YAML format within the SKILL.md file and does not include any standalone executable scripts for analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 05:18 AM