≈1k tokens on demand. Your agent reads this file only when SKILL.md points to it.
OWASP API Security Top 10 2023
Purpose: Use this file when API scope exists and you need a concise test strategy for API1-API10, including manual-vs-automated priority and Probe-specific emphasis.
Currency note (2026-05): OWASP API Security Top 10 2023 remains the active, published edition — there is no successor list yet (the next refresh cycle is in active community discussion at owasp.org/API-Security/). Treat the 2023 numbering as authoritative. For complementary AI/agent surfaces, pair with OWASP Top 10 for LLM Applications 2025 v2.0 (released 2024-11-18) and OWASP Top 10 for Agentic Applications (released 2025-12-09, see reference/llm-agent-security-2026.md).
BOLA prevalence update (2026-02): Wallarm's 2026 API ThreatStats Report found that 43% of all 2025 additions to CISA KEV (106 / 245) were API-related, 97% of API vulns are exploitable in a single request, and 52% of disclosed API breaches in 2025 traced to broken authentication. BOLA remains the #1 risk by attack volume.
Contents
- API1-API10 summary
- Test emphasis by risk
- Automation and manual split
- Probe quality gates
API1-API10 Summary
| Risk |
Name |
Severity |
Probe priority |
API1 |
Broken Object Level Authorization (BOLA) |
CRITICAL |
Highest |
API2 |
Broken Authentication |
CRITICAL |
Highest |
API3 |
Broken Object Property Level Authorization (BOPLA) |
HIGH |
High |
API4 |
Unrestricted Resource Consumption |
HIGH |
High |
API5 |
Broken Function Level Authorization (BFLA) |
HIGH |
High |
API6 |
Unrestricted Access to Sensitive Business Flows |
HIGH |
High |
API7 |
Server Side Request Forgery (SSRF) |
HIGH |
High |
API8 |
Security Misconfiguration |
MEDIUM-HIGH |
Medium |
API9 |
Improper Inventory Management |
MEDIUM |
Medium |
API10 |
Unsafe Consumption of APIs |
MEDIUM |
Medium |
Test Emphasis By Risk
| Risk |
What to test |
Automation |
Manual need |
API1 BOLA |
Object ID tampering, UUID predictability, ownership checks |
Partial |
Required |
API2 |
JWT handling, lockout, MFA gaps, reset flows |
Partial |
Required |
API3 BOPLA |
Mass assignment, excess response data |
Partial |
Required |
API4 |
Rate limits, payload size, pagination abuse |
Good |
Recommended |
API5 BFLA |
Admin endpoints, method switching, role escalation |
Partial |
Required |
API6 |
Coupon abuse, bots, scraping, business-flow abuse |
Weak |
Required |
API7 SSRF |
Internal URL fetch, metadata endpoints, redirect bypass |
Good |
Recommended |
API8 |
CORS, headers, defaults, debug exposure |
Good |
Recommended |
API9 |
Old versions, shadow APIs, deprecated assets |
Partial |
Required |
API10 |
Unsafe third-party API trust, schema mismatch, weak validation |
Weak |
Required |
Probe-Specific Defaults
- Always include
API1 / BOLA when API scope exists.
- Prioritize
API1, API2, API5, and API7 before medium-priority misconfiguration checks.
- Treat
API6 and API10 as business-context-heavy: automation is not enough.
- When the API exposes LLM-backed endpoints (chat, RAG, agent tool-call surfaces), cross-test with OWASP Top 10 for LLM 2025 v2.0 (LLM01 Prompt Injection, LLM02 Sensitive Information Disclosure, LLM06 Excessive Agency, LLM08 Vector and Embedding Weaknesses) and OWASP Top 10 for Agentic Applications 2025 (ASI01 Agent Goal Hijacking, indirect instruction injection via RAG content). Wallarm's 2026 API ThreatStats found 36% of AI vulnerabilities overlap with API vulnerabilities and MCP-related vulns grew 270% Q2-Q3 2025.
Quality Gates
| Condition |
Required action |
BOLA not tested |
Warn; this is the most common API attack family |
| Only automated API scanning used |
Add manual authorization and business-flow tests |
| Auth context missing |
Mark coverage as partial |
| No inventory/version review |
Warn for API9 blind spot |
