providers/crtsh_provider.py

@@ -3,7 +3,7 @@
 import json
 import re
 from pathlib import Path
-from typing import List, Dict, Any, Set
+from typing import List, Dict, Any, Set, Optional
 from urllib.parse import quote
 from datetime import datetime, timezone
 import requests
@@ -286,6 +286,17 @@ class CrtShProvider(BaseProvider):
             self.logger.logger.info(f"CrtSh processing cancelled before processing for domain: {query_domain}")
             return result
         
+        incompleteness_warning = self._check_for_incomplete_data(query_domain, certificates)
+        if incompleteness_warning:
+            result.add_attribute(
+                target_node=query_domain,
+                name="crtsh_data_warning",
+                value=incompleteness_warning,
+                attr_type='metadata',
+                provider=self.name,
+                confidence=1.0
+            )
+
         all_discovered_domains = set()
         processed_issuers = set()
 
@@ -578,3 +589,29 @@ class CrtShProvider(BaseProvider):
             return 'parent_domain'
         else:
             return 'related_domain'
+        
+    def _check_for_incomplete_data(self, domain: str, certificates: List[Dict[str, Any]]) -> Optional[str]:
+        """
+        Analyzes the certificate list to heuristically detect if the data from crt.sh is incomplete.
+        """
+        cert_count = len(certificates)
+
+        # Heuristic 1: Check if the number of certs hits a known hard limit.
+        if cert_count >= 10000:
+            return f"Result likely truncated; received {cert_count} certificates, which may be the maximum limit."
+
+        # Heuristic 2: Check if all returned certificates are old.
+        if cert_count > 1000: # Only apply this for a reasonable number of certs
+            latest_expiry = None
+            for cert in certificates:
+                try:
+                    not_after = self._parse_certificate_date(cert.get('not_after'))
+                    if latest_expiry is None or not_after > latest_expiry:
+                        latest_expiry = not_after
+                except (ValueError, TypeError):
+                    continue
+
+            if latest_expiry and (datetime.now(timezone.utc) - latest_expiry).days > 365:
+                 return f"Incomplete data suspected: The latest certificate expired more than a year ago ({latest_expiry.strftime('%Y-%m-%d')})."
+
+        return None

static/js/graph.js

@@ -1565,19 +1565,42 @@ class GraphManager {
     }
 
     /**
-     * Unhide all hidden nodes, excluding those within a large entity.
+     * FIXED: Unhide all hidden nodes, excluding large entity members and disconnected nodes.
+     * This prevents orphaned large entity members from appearing as free-floating nodes.
      */
     unhideAll() {
         const allHiddenNodes = this.nodes.get({
             filter: (node) => {
-                // Condition: Node is hidden AND it is NOT part of a large entity.
-                return node.hidden === true && !(node.metadata && node.metadata.large_entity_id);
+                // Skip nodes that are part of a large entity
+                if (node.metadata && node.metadata.large_entity_id) {
+                    return false;
+                }
+                
+                // Skip nodes that are not hidden
+                if (node.hidden !== true) {
+                    return false;
+                }
+                
+                // Skip nodes that have no edges (would appear disconnected)
+                const nodeId = node.id;
+                const hasIncomingEdges = this.edges.get().some(edge => edge.to === nodeId && !edge.hidden);
+                const hasOutgoingEdges = this.edges.get().some(edge => edge.from === nodeId && !edge.hidden);
+                
+                if (!hasIncomingEdges && !hasOutgoingEdges) {
+                    console.log(`Skipping disconnected node ${nodeId} from unhide`);
+                    return false;
+                }
+                
+                return true;
             }
         });
         
         if (allHiddenNodes.length > 0) {
+            console.log(`Unhiding ${allHiddenNodes.length} nodes with valid connections`);
             const updates = allHiddenNodes.map(node => ({ id: node.id, hidden: false }));
             this.nodes.update(updates);
+        } else {
+            console.log('No eligible nodes to unhide');
         }
     }
     

static/js/main.js

@@ -1397,11 +1397,32 @@ class DNSReconApp {
     }
 
     /**
-     * UPDATED: Generate details for standard nodes with organized attribute grouping
+     * UPDATED: Generate details for standard nodes with organized attribute grouping and data warnings
      */
     generateStandardNodeDetails(node) {
         let html = '';
 
+        // Check for and display a crt.sh data warning if it exists
+        const crtshWarningAttr = this.findAttributeByName(node.attributes, 'crtsh_data_warning');
+        if (crtshWarningAttr) {
+            html += `
+                <div class="modal-section" style="border-left: 3px solid #ff9900; background: rgba(255, 153, 0, 0.05);">
+                    <details open>
+                        <summary style="color: #ff9900;">
+                            <span>⚠️ Data Integrity Warning</span>
+                        </summary>
+                        <div class="modal-section-content">
+                            <p class="placeholder-subtext" style="color: #e0e0e0; font-size: 0.8rem; line-height: 1.5;">
+                                ${this.escapeHtml(crtshWarningAttr.value)}
+                                <br><br>
+                                This can occur for very large domains (e.g., google.com) where crt.sh may return a limited subset of all available certificates. As a result, the certificate status may not be fully representative.
+                            </p>
+                        </div>
+                    </details>
+                </div>
+            `;
+        }
+
         // Relationships sections
         html += this.generateRelationshipsSection(node);
 
@@ -1419,6 +1440,19 @@ class DNSReconApp {
         return html;
     }
 
+    /**
+     * Helper method to find an attribute by name in the standardized attributes list
+     * @param {Array} attributes - List of StandardAttribute objects
+     * @param {string} name - Attribute name to find
+     * @returns {Object|null} The attribute object if found, null otherwise
+     */
+    findAttributeByName(attributes, name) {
+        if (!Array.isArray(attributes)) {
+            return null;
+        }
+        return attributes.find(attr => attr.name === name) || null;
+    }
+
     generateOrganizedAttributesSection(attributes, nodeType) {
         if (!Array.isArray(attributes) || attributes.length === 0) {
             return '';