Skip to content

QWED's infrastructure is attested by Docker, Snyk, CircleCI, Netlify, Mintlify, Sentry, Cloudflare, CodeRabbit, NVIDIA Inception, Buildkite, GitLab, Heroku, Atlassian.

Back to Engine Stack
Data Flow Security

Trace Every Input
To Every Dangerous Sink.

AI-generated code can inadvertently pipe user input directly into dangerous functions. QWED's Taint Analyzer traces data flow paths from untrusted sources to sensitive sinks like eval(), exec(), and SQL queries.

Source (Tainted)
user_input = request.params["query"]
Propagation
processed = "SELECT * FROM users WHERE name = '" + user_input + "'"
Sink (Dangerous)
db.execute(processed)
BLOCKED: Tainted data reaches SQL execution sink without sanitization

What Taint Analysis Catches

AI code generators don't understand security context. They write functional code that often passes user input directly to dangerous operations.

Remote Code Execution

Detects when user-controlled data flows into eval(), exec(), subprocess, or system() calls without proper sanitization.

Cross-Site Scripting

Traces user input flowing into HTML templates, innerHTML assignments, or DOM manipulation without escaping.

SQL/NoSQL Injection

Identifies data paths from request parameters to database query construction without parameterized queries.

Security at the Data Flow Level

Static analysis finds dangerous functions. Taint analysis goes deeper—it proves whether untrusted data can actually reach those functions through the code's execution paths.

Beyond Pattern Matching

Traditional scanners use regex to find dangerous functions. QWED builds a complete data flow graph, tracking variable assignments, function returns, and object property access across the entire codebase.

Engine Capabilities

Source Identification
Automatically tag untrusted inputs (HTTP params, env vars, file reads).
Sink Detection
Comprehensive catalog of dangerous sinks (eval, exec, SQL, DOM, etc.).
Path Tracing
Full data flow graph from source to sink through variable assignments.
Sanitizer Recognition
Detect and validate sanitization functions in the data flow path.
Multi-Language Support
Python, JavaScript, TypeScript, Java, and Go analysis support.
Remediation Hints
Suggest specific sanitization functions for each detected vulnerability.

Ready to secure AI-generated code?