injection-agent

Blind injection vulnerabilities are validated using mandatory OOB DNS callbacks. The agent triggers dnslog.py to confirm out-of-band interactions for scenarios like blind SSRF, XXE, command injection, and SQLi where no direct response evidence exists.

You can test for SQLi, NoSQLi, XSS (stored/reflected/DOM), SSRF, XXE, SSTI, command injection, insecure deserialization, CRLF, XSLT, EL, JNDI, prototype pollution, type juggling, and HTTP request smuggling across both authenticated and unauthenticated scenarios.

Confirmed vulnerabilities require recorded HTTP interactions with complete headers and body. The agent enforces generating replayable http_test.py commands with complete request and response artifacts to satisfy strict evidence requirements.

Yes, injection testing covers both authenticated and unauthenticated scenarios. You provide available sessions and structured requests, and the agent models input-to-sink paths and generates targeted payloads for your authenticated API endpoints.

Recursive attack-surface expansion builds an initial endpoint and input inventory, then automatically expands to related parameters and deeper input contexts rather than stopping at the first anomaly, ensuring broader injection vulnerability coverage.

Payload generation systematically models input-to-sink paths to create targeted payloads for specific injection types like SSTI and prototype pollution. It validates these payloads by requiring replayable HTTP test commands and OOB DNS confirmation for blind cases.