gtm-closed-loss-reactivation
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill includes functionality to generate PDF artifacts using Python and the reportlab library.
- Evidence: The 'Artifact Generation' section in
SKILL.mdexplicitly lists 'Option C: PDF — Python + reportlab' for producing cheatsheets. - [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from CRM opportunity records and notes.
- Ingestion points: Step 1 of the
SKILL.mdworkflow extracts account history, loss notes, and conversation topics from a CRM. - Capability inventory: The skill is capable of generating email sequences, call talk tracks, and documents (PDF/HTML).
- Boundary markers: There are no explicit instructions to the agent to treat CRM notes as untrusted or to use specific delimiters.
- Sanitization: The skill does not describe any methods for cleaning or validating the contents of the ingested CRM records.
Audit Metadata