AI Security in Production — Part 1: Threat Modeling, Injection, Data & Identity Templates — 42 Templates & Checklists for Threat Modeling, Prompt-Injection Defense, Data Protection, Identity & Agent Tool Security

AI systems break every assumption classical security rests on — and this is the working defense for the first half of the discipline. 42 production templates & checklists distilled from Chapters 1–6 of AI Security in Production: the six-asset inventory, AI threat modeling end to end, six-layer prompt-injection defense with screening and ASR gates, data-as-payload protection, identity & access for AI workloads, and agent tool security priced by blast radius. PDF + Markdown source.

Your support bot just read the CEO's mail. Nothing failed.

No firewall failed. No credential was stolen. No malware ran. The system did what it was designed to do — follow instructions expressed in language — and that was the breach. AI Security in Production — Part 1 is the working defense for the half of AI security that decides whether that story happens to you: 42 copy-paste templates and checklists distilled from Chapters 1–6 of the dedicated security book — inventory the six assets you now defend, threat-model the system end to end, build the six-layer defense against prompt injection, govern the data because data is the payload, rebuild identity for workloads that act on a human's behalf, and price every agent action by its blast radius before you grant it.

Every template carries the book's tested positions built in: a one-page threat-model diagram with the context merge point drawn as the central boundary, STRIDE translated for AI and run per element, risk rated as blast radius × likelihood at production volume, a screening gateway whose worked shape downgrades sessions instead of only blocking them, attack success rate gates (the book's example programme: under 2% for tool-invocation goals, under 5% for prompt extraction), a four-class data scheme small enough to survive, retrieval that enforces source-system permissions in the query, two identities on every call with actions allowed only in their intersection, and agent tools classified READ / DRAFT / WRITE / TRANSACT with a policy file an auditor can read. Where the book gives an example — its gates, its worked policy-file limits — the template labels it as the book's example for you to tune. Bracketed fields are yours to fill; nothing here invents a threshold for your system.

What's inside — 42 templates & checklists in 7 sections

1. Why AI Breaks Classical Security & the Asset Inventory (5) — Six-Asset AI System Inventory · Four Broken Assumptions Audit · End-to-End Attack Surface & Trust-Boundary Walk · Adversary-to-Control Mapping Sheet · Four Incident Patterns Business-Case Worksheet — Leak, Lie, Act, Bill

2. Threat Modeling for AI Systems, End to End (6) — One-Page AI Data-Flow Diagram Specification · STRIDE-for-AI Enumeration Worksheet · Blast-Radius & Likelihood-at-Volume Rating Sheet · AI Threat Register Template · Threat-Modeling Workshop Kit — Agenda, Triggers & Register-to-Backlog · House Threat Catalogue & Threat-Model Anti-Pattern Review

3. Prompt Injection — Defense in Depth for a Persuadable Component (6) — Injection Species Triage Sheet — Direct, Indirect, Jailbreak, Exfiltration Chain · Six-Layer Injection Defense Architecture Checklist · Screening Gateway Specification — Rules, Classifier & the DOWNGRADE Decision · Payload Pattern Recognition Sheet — Five Patterns Worth Knowing by Sight · Attack Success Rate Corpus & Release-Gate Sheet · Session Downgrade Ladder — L0 to L4

4. Data Is the Payload — Protecting Information in the Age of Prompts (6) — Four-Class Data Classification Scheme · Ingestion Redaction Policy — Mask, Tokenise, Remove · Retrieval Authorization Enforcement Checklist — the Model Is Not a Permissions Bypass · Logs, Retention, Deletion & Vector-Store Protection Checklist · Prompt Data-Minimisation Checklist — the Editor's Discipline · Data-Subject Rights & Training-Data Worksheet

5. Identity & Access for AI Workloads (6) — Two-Identity Intersection Worksheet · Standing-Key Elimination & Secrets Hygiene Checklist · AI Gateway Control Checklist — One Door, Every Control · Four-Level Authorization Policy Worksheet · Prompts-as-Code Change-Control Checklist · Just-in-Time Access, Non-Human Identity Registry & the Two Doors Sheet

6. Agents, Tools & Autonomy — Securing Actions by Blast Radius (7) — Autonomy Tier Classification Worksheet — READ / DRAFT / WRITE / TRANSACT · Tool Permission Policy File Template — Capabilities, Not Keys · Sandbox & Execution Control Checklist · Human Approval Design Checklist — Approval That Actually Works · Agent Budgets, Loop Detection & Circuit-Breaker Configuration — Enforced Outside the Agent · Multi-Agent Delegation Security Sheet — Govern the Graph, Not Just the Nodes · Agent Lifecycle Sheet — Permission Audits, Reversibility & the Off-Ramp

7. Part 1 Gates & Working Artefacts (6) — Trust Boundary Inventory Sheet · Attacker Recon Conversation Exercise · Screening Operations Tuning Log · Classification Review Board & Corpus Census Sheet · Authorization Review & Access-Request Sheet · Part 1 Production Security Gate — Chapters 1–6 Sign-Off — the book's six chapter checklists consolidated into one evidence-backed gate (35 lines)

Who this is for

CISOs making their first-year AI decisions, security architects asked to approve designs they did not create, and platform and AI engineering leads who own systems that must survive contact with real users and real attackers. The source book assumes you know how enterprise systems are generally built and defended; no machine-learning expertise is required. If nobody in your organisation can answer what your flagship AI system can read, what it can do, and whose instructions it will follow — in one sitting — start at Template 1.

What this pack is not

Not Part 2. Model supply chain and provenance, guardrails as an engineered control system, the red-team program, monitoring / detection / the AI incident runbook, the framework compliance map, and the defense-program roadmap are the book's Chapters 7–12. They live in the Part 2 companion pack; nothing here restates them, and this pack is complete and useful without them.

No compliance crosswalk or certification. Threat modeling here enumerates with MITRE ATLAS and the OWASP LLM Top 10, as the book's Chapter 2 prescribes. No NIST AI RMF, ISO/IEC 42001, or EU AI Act mapping ships in this pack, and the Part 1 gate in Template 42 is a Chapters 1–6 production gate — not a compliance certification.

No runnable code or tooling. The book's gateway, policy, and redaction sketches are simplified, structurally honest shapes. The pack ships the templates that specify, audit, and operate these components — not a working gateway, classifier, or policy engine.

No vendor or product guidance. The gateway is a pattern — built or bought; the pattern matters more than the product, in the book's words. No security product is selected, installed, or configured here.

No guarantee of security. The book's own position: no checklist makes a language-model system secure. These templates manage, measure, and bound risk — injection defense is tracked by attack success rate precisely because it is reduced, never eliminated.

Not the cross-book security slice. The companion AI Security Templates & Checklists pack holds the MCP-and-enterprise-agents security artifacts (OAuth deployment, MCP attack surface, tool-poisoning change control). This pack is the dedicated book's deeper method — threat modeling, injection defense-in-depth, data-as-payload protection, identity for AI workloads, and tool / action / autonomy security. The two packs do not restate each other.

Companion products

The AI Security Templates & Checklists pack is the cross-book companion — the MCP and enterprise-agent security slice this pack deliberately does not restate. AI Security in Production — Part 2: Supply Chain, Guardrails, Red Team & the Defense Program Templates is this pack's sibling, built from Chapters 7–12 of the same book: a companion for when you are ready for the second half — never a requirement for this pack to work. This pack stands alone.

Note: the $99 Architecture Templates Bundle currently contains Products #1–#3 only. This pack launches standalone; no bundle inclusion is promised for this product.

Built by the author of AI Security in Production: The Enterprise Defense Handbook, from the defense discipline behind production enterprise AI systems.

About the author

Naveen Kumar Josyula is the author of 35 AI books on Amazon KDP and the CEO and founder of AISoltrics Technologies Inc. He is the author of the source book, AI Security in Production: The Enterprise Defense Handbook, from which every template in this pack is distilled.

Support promise

If anything in this file is wrong or unclear, email support@aisoltrics.com before you judge it — we fix verified issues within one business day and send you the corrected file free. (You can also reach the founder's office at aadhya@aisoltrics.com.) Updates to this product are free to past buyers via your Payhip library. Sold by AISoltrics Technologies Inc.