Centralized catalog of standard risks for quick instantiation
910 templates found
| Code | Name | Category | Status | Actions |
|---|---|---|---|---|
| AI-01 | Prompt Injection Data Leakage Prompt injection causes disclosure of sensitive data. | AI / GenAI & Model Risks (AI) | Active | |
| AI-02 | Sensitive Data in Training/Context Confidential data included in training/context causes leakage. | AI / GenAI & Model Risks (AI) | Active | |
| AI-03 | Inaccurate Compliance Advice Model errors produce wrong control/risk guidance. | AI / GenAI & Model Risks (AI) | Active | |
| AI-04 | No Audit Trail for AI Decisions AI actions lack audit trail, reducing accountability. | AI / GenAI & Model Risks (AI) | Active | |
| AI-05 | Weak AI Access Controls AI tools accessible broadly enable misuse. | AI / GenAI & Model Risks (AI) | Active | |
| AI-06 | Tool/Plugin Abuse via Agent Agent tools abused to perform unauthorized actions. | AI / GenAI & Model Risks (AI) | Active | |
| AI-07 | External LLM Policy Violation Sensitive data sent to external LLM violates policy/law. | AI / GenAI & Model Risks (AI) | Active | |
| AI-08 | Model Drift in Scoring Model drift changes risk scoring unpredictably. | AI / GenAI & Model Risks (AI) | Active | |
| AI-09 | Hallucinated Control Mappings Hallucinations create incorrect control mappings and plans. | AI / GenAI & Model Risks (AI) | Active | |
| AI-10 | No AI Red-Teaming Lack of testing leaves exploitable behaviors. | AI / GenAI & Model Risks (AI) | Active | |
| AI-11 | Prompt Logging Contains Secrets Prompt/response logs contain secrets and PII. | AI / GenAI & Model Risks (AI) | Active | |
| AI-12 | No Output Validation AI outputs not validated leading to unsafe actions. | AI / GenAI & Model Risks (AI) | Active | |
| AI-13 | Weak Prompt Template Governance Prompts changed without review; behavior shifts. | AI / GenAI & Model Risks (AI) | Active | |
| AI-14 | No Guardrails for Sensitive Topics AI provides unsafe guidance or discloses policies. | AI / GenAI & Model Risks (AI) | Active | |
| AI-15 | Training Data Poisoning Risk Poisoned data influences outputs and decisions. | AI / GenAI & Model Risks (AI) | Active | |
| AI-16 | RAG Retrieval Overexposes Documents Retrieval shows docs user should not access. | AI / GenAI & Model Risks (AI) | Active | |
| AI-17 | No Tenant Isolation in AI Context Cross-tenant context leakage via embeddings/prompts. | AI / GenAI & Model Risks (AI) | Active | |
| AI-18 | Embedding Store Access Too Broad Vector store accessible broadly; data leakage. | AI / GenAI & Model Risks (AI) | Active | |
| AI-19 | No Citation/Provenance Answers lack sources; trust decreases. | AI / GenAI & Model Risks (AI) | Active | |
| AI-20 | No Human-in-the-Loop for High-Risk Actions AI executes high-risk actions without approval. | AI / GenAI & Model Risks (AI) | Active | |
| AI-21 | Over-Reliance on AI for Risk Ratings Humans accept AI ratings without review. | AI / GenAI & Model Risks (AI) | Active | |
| AI-22 | LLM Vendor Outage LLM provider outage impacts operations. | AI / GenAI & Model Risks (AI) | Active | |
| AI-23 | Cost Explosion Uncontrolled usage causes unexpected costs. | AI / GenAI & Model Risks (AI) | Active | |
| AI-24 | Prompt/Model Versioning Missing Cannot reproduce decisions; audit fails. | AI / GenAI & Model Risks (AI) | Active | |
| AI-25 | Insecure Tool Credentials Tool credentials used by agents stored insecurely. | AI / GenAI & Model Risks (AI) | Active | |
| AI-26 | PII Sent to Tools PII sent to tools without minimization. | AI / GenAI & Model Risks (AI) | Active | |
| AI-27 | Weak Content Filtering Malicious content influences agent behavior. | AI / GenAI & Model Risks (AI) | Active | |
| AI-28 | No Monitoring for AI Abuse Abuse patterns not detected. | AI / GenAI & Model Risks (AI) | Active | |
| AI-29 | No Model Access Segmentation All users use same powerful model; misuse risk. | AI / GenAI & Model Risks (AI) | Active | |
| AI-30 | Unclear Data Retention by Vendor Vendor retention unknown; compliance risk. | AI / GenAI & Model Risks (AI) | Active | |
| AI-31 | Inadequate Bias Testing Bias leads to unfair recommendations. | AI / GenAI & Model Risks (AI) | Active | |
| AI-32 | No Incident Response for AI AI-specific incident playbooks missing. | AI / GenAI & Model Risks (AI) | Active | |
| AI-33 | Insecure Fine-Tuning Process Fine-tuning leaks data or introduces vulnerabilities. | AI / GenAI & Model Risks (AI) | Active | |
| AI-34 | Prompt Injection into Tools Injected prompts cause tools to act maliciously. | AI / GenAI & Model Risks (AI) | Active | |
| AI-35 | No Data Classification in AI Pipelines AI pipelines ignore classification; leaks occur. | AI / GenAI & Model Risks (AI) | Active | |
| API-01 | Missing API Authentication Endpoints lack authentication, enabling unauthorized access. | API Security (API) | Active | |
| API-02 | Weak JWT Validation Improper token validation enables token forgery. | API Security (API) | Active | |
| API-03 | Excessive Data Exposure APIs return more data than needed, exposing sensitive fields. | API Security (API) | Active | |
| API-04 | No Schema/Input Validation Invalid input leads to injection and abuse. | API Security (API) | Active | |
| API-05 | Insecure CORS Policy Overly permissive CORS enables cross-site abuse. | API Security (API) | Active | |
| API-06 | No API Throttling No throttling allows DoS and brute force. | API Security (API) | Active | |
| API-07 | Broken Object-Level Authorization BOLA enables attackers to access others’ records. | API Security (API) | Active | |
| API-08 | API Keys Embedded in Clients Keys in apps are stolen and abused. | API Security (API) | Active | |
| API-09 | No API Audit Logging Missing logs prevent investigation and accountability. | API Security (API) | Active | |
| API-10 | Deprecated APIs Left Active Old endpoints remain exploitable and unpatched. | API Security (API) | Active | |
| API-11 | Broken Function-Level Authorization Users can call admin functions via crafted requests. | API Security (API) | Active | |
| API-12 | Mass Assignment Clients set protected fields via payload manipulation. | API Security (API) | Active | |
| API-13 | No Pagination Limits Large responses enable DoS and data scraping. | API Security (API) | Active | |
| API-14 | Improper Error Handling Errors reveal internal details and stack traces. | API Security (API) | Active | |
| API-15 | Insecure Webhook Endpoints Webhooks lack verification allowing spoofed calls. | API Security (API) | Active |
