Deal Slippage Diagnosis

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill utilizes standard platform actions (get_pipeline_deals, get_deal, get_deal_activities) to perform data operations, ensuring actions are contained within the agent's authorized capability scope.\n- [SAFE]: No evidence of hardcoded credentials, obfuscated instructions, or unauthorized network communication to external domains was found.\n- [PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection due to its analysis of untrusted external data.\n
  • Ingestion points: The skill fetches and analyzes CRM activity notes and meeting transcripts to detect competitive mentions and other risk signals (referenced in Signal 14 of slippage-signals.md).\n
  • Boundary markers: Explicit delimiters or 'ignore' instructions for the interpolated CRM data are not specified in the skill logic.\n
  • Capability inventory: The skill has crm (read) and tasks (preview/create) capabilities, allowing it to generate prioritized rescue plans and task previews based on analyzed data.\n
  • Sanitization: There is no documented mechanism for sanitizing or filtering instructions that might be embedded within the CRM notes or transcripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 11:37 AM