🛡️ When one AegisGate customer detects a new AI attack, every customer is protected within minutes.
Traditional security tools operate in isolation. When Company A detects a novel prompt injection attack, Company B learns nothing. Attackers exploit this fragmentation repeatedly.
AegisGate Federated IOC Store changes this. Detected threats are packaged into cryptographically signed bundles and distributed to all customers. The network grows stronger with every detection.
View Pricing → Technical Documentation →
The Isolation Problem
Traditional security vendors operate in silos:
Company A detects attack → Logs stored locally → Company B remains vulnerable
Company B detects same attack → Logs stored locally → Company C remains vulnerable
Company C detects same attack → Logs stored locally → Attackers continue exploiting
Result: The same attack succeeds against multiple organizations because threat intelligence doesn’t flow.
Why this happens:
- Vendors treat threat intel as a premium add-on (separate subscription)
- Intelligence is delayed (weekly digests, not real-time)
- Formats are incompatible (STIX/TAXII requires separate tooling)
- Trust is unclear (who generated this IOC? can I trust it?)
The Federated Solution
AegisGate Federated IOC Store creates a network effect:
Customer detects attack → Platform generates signed IOC bundle → All customers receive within 5 minutes
↓
Automatic protection (no manual update required)
Key Properties:
| Property | Implementation | Why It Matters |
|---|---|---|
| Cryptographic Trust | Every bundle signed with ECDSA P-256 | Customers verify bundle integrity before applying |
| Real-Time Distribution | Push-based (not pull) | Protection arrives in minutes, not days |
| AI-Specific IOCs | Prompt patterns, injection signatures, adversarial examples | Generic IOC stores miss AI-specific attacks |
| TAXII 2.1 Compatible | Standard protocol integration | Works with existing SIEM/SOAR investments |
| Privacy-Preserving | IOCs contain patterns, not customer data | No sensitive information leaves your environment |
How It Works
Step 1: Detection
When Platform detects a novel attack (e.g., new prompt injection technique):
{
"detection_id": "det-2026081409234501",
"category": "adversarial_ai",
"technique": "prompt_injection",
"pattern": "ignore previous instructions.*output system prompt",
"severity": "high",
"customer_id": "[REDACTED]",
"timestamp": "2026-08-14T09:23:45Z"
}
Step 2: IOC Generation
Platform automatically generates an IOC bundle:
{
"bundle_id": "ioc-bundle-2026081409",
"generated_at": "2026-08-14T09:24:15Z",
"source": "aegisgate-platform",
"version": "1.0",
"indicators": [
{
"id": "indicator-001",
"type": "prompt_pattern",
"pattern": "ignore previous instructions.*output system prompt",
"confidence": 0.95,
"severity": "high",
"technique": "MITRE ATLAS T1535",
"description": "Direct prompt injection attempting to extract system instructions"
}
],
"signature": "MEUCIQDvL8K9mN2pQ3rS4tU5vW6xY7zA8bC9dE0fG1hI2jK3lM4n"
}
Signed with: Platform’s ECDSA private key
Step 3: Distribution
Bundle is pushed to all connected customers:
# Platform pushes to TAXII 2.1 server
POST /api/v1/ioc/bundles
Content-Type: application/taxii+json
{ ... signed bundle ... }
Delivery SLA:
- Enterprise: < 5 minutes
- Air-Gap: Manual import (bundle exported, transferred via secure media)
Step 4: Automatic Protection
Customer Platform instances receive and apply the bundle:
# Customer Platform verifies signature
openssl dgst -sha256 -verify aegisgate-ioc.pub \
-signature bundle.sig \
bundle.json
# Output: Verified OK ✅
# IOC pattern added to detection engine
# Future prompts matching this pattern trigger immediate block
No manual intervention required.
IOC Types
Federated IOC Store tracks AI-specific indicators:
| IOC Type | Example | Detection Use Case |
|---|---|---|
| Prompt Patterns | Regex signatures for injection techniques | Block known adversarial prompts |
| Response Signatures | Patterns indicating model compromise | Detect when AI output has been manipulated |
| Behavioral Anomalies | Unusual token patterns, encoding tricks | Identify obfuscated attacks |
| Adversarial Examples | Specific input sequences that bypass safety | Block evasion techniques |
| Extraction Templates | Known system prompt extraction patterns | Prevent model intellectual property theft |
Not included: Generic malware hashes, IP addresses, domains (these are handled by traditional IOC feeds)
Integration with Existing Tools
TAXII 2.1 Integration
Federated IOC Store speaks TAXII 2.1 (Trusted Automated Exchange of Intelligence Information):
# Subscribe to AegisGate TAXII server
GET /taxii2.1/collections/aegisgate-ioc
Accept: application/taxii+json
Authorization: Bearer <api_key>
Compatible with:
- Splunk Enterprise Security
- IBM QRadar
- Microsoft Sentinel
- Palo Alto Cortex XSOAR
- Any TAXII 2.1 client
STIX 2.1 Packaging
IOCs are packaged in STIX 2.1 (Structured Threat Information Expression):
{
"type": "indicator",
"spec_version": "2.1",
"id": "indicator--aegisgate-001",
"pattern": "[process:command_line MATCHES 'ignore previous instructions.*output system prompt']",
"pattern_type": "regex",
"valid_from": "2026-08-14T09:24:15Z",
"confidence": 95,
"severity": "high",
"labels": ["adversarial-ai", "prompt-injection", "mitre-atlas-t1535"]
}
Why STIX matters:
- Standard format understood by SIEM/SOAR tools
- No custom parsers required
- Integrates with existing threat intelligence workflows
Privacy & Data Protection
What IS shared:
- IOC patterns (regex, signatures)
- Attack technique metadata (MITRE ATLAS mapping)
- Severity and confidence scores
- Timestamp (when detected)
What is NEVER shared:
- Customer identity (anonymized)
- Actual prompts or responses (only patterns)
- Sensitive data (PII, secrets, proprietary code)
- Internal system details
Technical Guarantee:
- IOCs are generated from detection patterns, not raw data
- Customer IDs are redacted before bundle creation
- Bundles are signed — tampering is detectable
Use Cases
Financial Services
Problem: Coordinated attacks targeting banking AI assistants (prompt injection to bypass transaction limits).
Federated IOC Solution:
- Customer A detects attack → IOC bundle generated
- Customers B, C, D receive bundle within 5 minutes
- All customers protected before attackers can scale
Compliance Mapping:
- FFIEC CAT: TI.A.1 (Threat intelligence received)
- GLBA: Safeguards Rule (information security)
Healthcare
Problem: Attackers targeting patient portal AI chatbots to extract PHI.
Federated IOC Solution:
- Hospital system A detects PHI extraction attempt
- IOC pattern shared with all healthcare customers
- Pattern added to HIPAA compliance reporting
Compliance Mapping:
- HIPAA Security Rule: §164.308(a)(1)(ii)(B) (Risk analysis)
- HITECH Act: Breach prevention
Government
Problem: Adversarial actors attempting to extract classified information via AI systems.
Federated IOC Solution:
- Agency A detects classified data extraction attempt
- IOC shared with cleared government customers only (segmented feed)
- Pattern integrated with CISA reporting
Compliance Mapping:
- NIST SP 800-53: SI-4 (Information system monitoring)
- FedRAMP: IR-4 (Incident handling)
Technology Sector
Problem: Competitors attempting to extract proprietary model weights or training data.
Federated IOC Solution:
- Tech company A detects model extraction attack
- IOC pattern shared with all tech sector customers
- Attack technique added to threat intelligence briefing
Compliance Mapping:
- SOC 2: CC6.1 (Logical and physical access controls)
- ISO 27001: A.12.4 (Logging and monitoring)
API Reference
Subscribe to IOC Feed
GET /api/v1/ioc/subscribe
Authorization: Bearer <api_key>
Response:
{
"subscription_id": "sub-ioc-001",
"feed_url": "https://ioc.aegisgate.io/taxii2.1/collections/aegisgate-ioc",
"api_key": "<unique_key>",
"created_at": "2026-08-14T09:00:00Z"
}
Get Latest Bundles
GET /api/v1/ioc/bundles?limit=10
Authorization: Bearer <api_key>
Response:
{
"bundles": [
{
"bundle_id": "ioc-bundle-2026081409",
"generated_at": "2026-08-14T09:24:15Z",
"indicator_count": 3,
"severity": "high",
"verified": true
}
]
}
Export IOC Bundle (STIX Format)
GET /api/v1/ioc/bundles/{bundle_id}/export?format=stix
Authorization: Bearer <api_key>
Accept: application/stix+json
Response: STIX 2.1 formatted bundle (ready for SIEM import)
Manual Import (Air-Gap)
# Export bundle to file
POST /api/v1/ioc/bundles/{bundle_id}/export
Content-Type: application/json
{
"destination": "file",
"path": "/tmp/ioc-bundle-2026081409.json"
}
# Transfer via secure media to air-gapped environment
# Import on air-gapped Platform instance
POST /api/v1/ioc/import
Content-Type: application/json
{
"bundle_file": "/media/secure/ioc-bundle-2026081409.json"
}
Tier & Availability
| Tier | Federated IOC Store Features |
|---|---|
| Community | ❌ Not available |
| Developer | ❌ Not available |
| Professional | ✅ Real-time IOC feed, TAXII 2.1 integration, STIX 2.1 export |
| Enterprise | ✅ All Professional features + manual bundle import for air-gap, segmented feeds, redistribution rights |
Minimum Version: Platform v4.2.0+
Network Effect: The Moat
Traditional Security: Every customer starts from zero. Threat intelligence is siloed.
AegisGate Federated IOC: Every customer benefits from every other customer’s detections.
1 customer detects attack → 100 customers protected
10 customers detect attacks → 1,000 customers protected
100 customers detect attacks → 10,000 customers protected
Competitive Advantage:
- Competitors cannot replicate network effects (requires installed base)
- Value increases with every customer (positive feedback loop)
- Switching costs increase over time (historical IOC data retained)
Competitive Comparison
| Feature | AegisGate | Competitor A | Competitor B | Competitor C |
|---|---|---|---|---|
| AI-specific IOCs | ✅ Yes (prompt patterns, adversarial examples) | ❌ Generic malware only | ⚠️ Limited (manual curation) | ❌ No |
| Real-time distribution | ✅ < 5 minutes | ❌ Weekly digests | ⚠️ Hourly | ❌ N/A |
| Cryptographic signing | ✅ ECDSA P-256 | ❌ Unsigned | ❌ Unsigned | ❌ N/A |
| TAXII 2.1 integration | ✅ Yes | ⚠️ Enterprise add-on | ❌ No | ❌ N/A |
| STIX 2.1 format | ✅ Yes | ⚠️ STIX 1.0 only | ❌ Proprietary | ❌ N/A |
| Privacy-preserving | ✅ Patterns only (no raw data) | ❌ Raw logs shared | ⚠️ Anonymized (re-identifiable) | ❌ N/A |
| Air-gap support | ✅ Manual bundle import | ❌ Cloud-only | ❌ Cloud-only | ❌ N/A |
| Network effect | ✅ Yes (crowdsourced) | ❌ Vendor-curated only | ❌ Vendor-curated only | ❌ N/A |
Bottom line: Federated IOC Store is the only AI-specific threat intelligence network with real-time distribution and cryptographic trust.
Frequently Asked Questions
Can I opt out of sharing IOCs?
Yes. Enterprise customers can choose to receive IOCs without contributing. However, we encourage participation — the network effect only works when customers contribute detections.
What if an IOC generates false positives in my environment?
IOCs include confidence scores. Low-confidence IOCs (< 0.7) are flagged for review before automatic blocking. You can also create allowlists for specific patterns.
How do I verify an IOC bundle's signature?
Bundles include an ECDSA P-256 signature. Use OpenSSL or any cryptographic library to verify: openssl dgst -sha256 -verify aegisgate-ioc.pub -signature bundle.sig bundle.json
Can I share IOCs with my own customers?
Yes. Enterprise tier includes redistribution rights. You can integrate AegisGate IOCs into your own threat intelligence platform or MSSP offerings.
What happens if the IOC feed is unavailable?
Platform continues operating with locally cached IOCs. When connectivity is restored, missed bundles are backfilled automatically.
Ready to Join the Network?
Federated IOC Store is available in Professional tier and above.
Start Free Trial → Technical Documentation → Contact Sales →