Threat model walkthrough: prompt injection, exfiltration, poisoning
A checklist of threat model, data governance, prompt-injection defense, audit trail, vendor review, and incident-response questions. It is not a certification and not a record of audits we performed.
What's inside
Sections included in this template
Data governance controls: retention, redaction, purpose limitation
Authentication, authorization, and per-inference logging
Guardrails and output validation requirements
Model card and evaluation documentation requirements
Vendor risk review: OpenAI, Anthropic, Bedrock, Azure OpenAI
SOC 2, HIPAA, and GDPR cross-reference matrix
Incident response playbook with disclosure timelines
Evidence artifact templates for each control
Download
Get the full editable template
The file is not published on this site. Request a copy and we will follow up.