Apigee as a API fuzzer

An API fuzzer sends a large number of malformed, unexpected, or random inputs to an endpoint to uncover crashes, unhandled exceptions, and inconsistent behavior. Its purpose is to surface instability and error paths rather than to prove exploitability. It does not perform intrusive attacks such as active SQL injection or command injection, which require precise payload crafting and stateful interactions outside its scope.

The tool does not detect business logic flaws that depend on domain-specific workflows or context. It also does not identify blind SSRF that requires out-of-band confirmation. Because it does not execute code or maintain authenticated application state, it cannot replace a human pentester for high-stakes audits.

Findings align with three frameworks by mapping directly to their controls: PCI-DSS 4.0, SOC 2 Type II, and OWASP API Top 10 (2023). Each reported issue includes guidance on how it maps to relevant requirements or controls within these frameworks.

For other standards, middleBrick supports audit evidence collection and helps you prepare for security reviews. The tool surfaces findings relevant to controls described in HIPAA, GDPR, ISO 27001, NIST, CCPA, and similar regulations using an alignment approach rather than compliance claims.

As a black-box scanner, the tool probes endpoints without agents, SDKs, or code access. It sends only read-only methods (GET and HEAD) plus text-only POST for LLM probes, avoiding destructive payloads. Scan completion typically occurs in under a minute.

The engine tests input validation by injecting unexpected types, boundary values, and encoding variants. It checks for CORS wildcard usage with and without credentials, dangerous HTTP methods, debug endpoints, and oversized or paginated responses. All network tests are blocked at multiple layers for private IPs, localhost, and cloud metadata endpoints.

Authenticated scanning is available from the Starter tier onward, supporting Bearer tokens, API keys, Basic auth, and cookies. Domain verification is enforced through a DNS TXT record or an HTTP well-known file to ensure only the domain owner can submit credentials.

The scanner forwards a restricted allowlist of headers, including Authorization, X-API-Key, Cookie, and X-Custom-*. It never modifies, deletes, or writes data. Customer data is deletable on demand and purged within 30 days of cancellation, and it is never used for model training.

The scanner includes 18 adversarial probes for LLM / AI Security across three scan tiers: Quick, Standard, and Deep. These probes test system prompt extraction, instruction override, DAN and roleplay jailbreaks, data exfiltration attempts, cost exploitation, base64 and ROT13 encoding bypasses, translation-embedded injection, few-shot poisoning, markdown injection, multi-turn manipulation, indirect prompt injection, token smuggling, tool abuse, nested instruction injection, and PII extraction.

For API definitions, the tool parses OpenAPI 3.0, 3.1, and Swagger 2.0 with recursive $ref resolution. It cross-references spec definitions against runtime findings to highlight undefined security schemes, sensitive fields, deprecated operations, and missing pagination.