Zero-Day Vulnerability Detector

Filters & advanced settings

File extension filter (empty = all)

Comma-separated extensions. Example: .php, .inc to only scan PHP files.

Skip directories (extra exclusions)

Comma-separated directory names added to the built-in exclusions.

Tip. Filtering by extension speeds up analysis. Common: .php · .js · .py · .java · .rb.

Use true positives only Only include examples that were confirmed as true positives by the AI judge.

WordPress plugin mode Enable endpoint reachability analysis. Only show vulnerabilities in functions reachable from WordPress REST API, AJAX handlers, or admin pages.

WordPress mode. Uses WPTraceLib to detect WordPress endpoints (REST API, AJAX, Admin pages) and traces call chains from callbacks to determine which vulnerable functions are actually reachable from external input.
7-level auth hierarchy: Unauthenticated → Subscriber → Contributor → Author → Editor → Admin → SuperAdmin.

Context analysis depth 3

1 (direct only) 2 3 (default) 4 5 (deep)

Controls how deep to recursively fetch function definitions. Higher depth provides more context but takes longer.

Rule selection & AI options

Rule-based filtering (optional). When rules are selected, two-pass detection is used:

1. Pattern matching. Extract lines matching security-sensitive patterns.
2. Initial pass. AI performs quick triage on matched patterns.
3. Deep scan. Full analysis of files flagged in the initial pass.

Without rules. All source files will be analyzed directly (slower but more comprehensive).

Loading available rules...

Select rules for two-pass detection.

✓ Select all ⌫ Deselect all

Enable AI analysis Run AI analysis on sink matches. Disable to only run pattern matching + InputTracer filtering.

Enable InputTracer filtering Only analyze sink matches where variables receive user input. Reduces false positives by filtering out matches without tainted data.

InputTracer. Performs backward taint analysis on each variable in sink-matched lines to determine if they trace back to user input sources (e.g., $_GET, $_POST, req.body).
Supported languages: PHP, JavaScript.

∅

No rules found for selected CWEs

Rules are built-in for common vulnerability types. View Rules Library

Example selection

Choose how many examples to use and optionally select specific examples or use AI to identify unique examples with different fix patterns.

Example count

Number of examples to use per CWE (default: 5).

AI unique examples ⚲ Identify unique examples (AI)

Click to let AI identify unique examples with different fix patterns.

AI is analyzing examples to find unique patterns...

AI identified unique examples. Select the ones you want to use below.

---

Manually select examples Or browse all available examples and choose manually.

Loading available examples...

Select examples to include (or leave unchecked to use auto-selection).

∅

No examples found

▶ Start 0day Detection

Analysis options

Context analysis depth 3

1 (direct) 2 3 (default) 4 5 (deep)

Enable AI analysis Run AI vulnerability analysis on matched patterns.

Enable InputTracer filtering Only analyze matches where variables receive user input.

Use grep / sink rules Pre-filter files using pattern matching before AI analysis (faster for large codebases).

Loading available rules...

No rules available for selected CWEs

✓ Select all ⌫ Deselect all