Compliance

How AgentWeave SDK supports compliance with SOC 2, HIPAA, PCI DSS, GDPR, and FedRAMP

Edit on GitHub

Compliance Considerations

AgentWeave SDK provides security controls and features that help organizations meet various compliance requirements. This guide explains how AgentWeave addresses common compliance frameworks.

Important: AgentWeave provides technical controls, but compliance requires organizational policies, procedures, and audits. Consult with compliance experts for your specific requirements.

Table of Contents

  1. Compliance Considerations
    1. How AgentWeave Helps with Compliance
    2. SOC 2 (System and Organization Controls)
      1. Overview
      2. Relevant Trust Service Criteria
      3. SOC 2 Evidence Collection
    3. HIPAA (Health Insurance Portability and Accountability Act)
      1. Overview
      2. Relevant Safeguards
      3. HIPAA Implementation Recommendations
    4. PCI DSS (Payment Card Industry Data Security Standard)
      1. Overview
      2. Relevant Requirements
    5. GDPR (General Data Protection Regulation)
      1. Overview
      2. Relevant Articles
      3. GDPR Implementation Recommendations
    6. FedRAMP (Federal Risk and Authorization Management Program)
      1. Overview
      2. Relevant Controls (NIST SP 800-53)
      3. FedRAMP Implementation
    7. Compliance Checklist
      1. Authentication & Identity
      2. Authorization
      3. Encryption
      4. Audit Logging
      5. Monitoring
      6. Incident Response
      7. Documentation
    8. Evidence for Auditors
      1. Technical Evidence
      2. Process Evidence
    9. Important Notes
    10. Next Steps
    11. References

How AgentWeave Helps with Compliance

AgentWeave provides several security controls that map to common compliance requirements:

Control AgentWeave Feature
Authentication SPIFFE/SPIRE cryptographic identity
Authorization OPA policy-based access control
Encryption in Transit TLS 1.3 with mTLS
Audit Logging Comprehensive request/decision logging
Access Control Default deny, least privilege
Identity Management Automatic credential rotation
Monitoring Metrics, traces, logs (observability)
Integrity Cryptographic signatures, immutable logs

SOC 2 (System and Organization Controls)

Overview

SOC 2 focuses on controls relevant to security, availability, processing integrity, confidentiality, and privacy of customer data.

Relevant Trust Service Criteria

CC6.1 - Logical and Physical Access Controls

Requirement: The entity implements logical access security controls to protect system resources from unauthorized access.

AgentWeave Controls:

  1. Cryptographic Identity (CC6.1)
    • Every agent has SPIFFE identity (cannot be forged)
    • No shared credentials or API keys
    • Automatic rotation prevents credential compromise
1
2
3
4

# Configuration evidence
identity:
  provider: "spire"
  socket_path: "/run/spire/sockets/agent.sock"
  1. Mutual Authentication (CC6.1)
    • mTLS required for all agent communication
    • Both parties verify each other's identity
    • Cannot bypass authentication
  2. Default Deny Authorization (CC6.1)
    • All requests denied unless explicitly allowed
    • Policy-based access control
    • Centralized policy management
1
2

# Policy evidence
default allow := false  # Must explicitly grant access

CC6.2 - System Access is Removed When Access is No Longer Required

AgentWeave Controls:

  1. Short-lived Credentials (CC6.2)
    • SVIDs expire after 1 hour by default
    • Automatic rotation at 50% TTL
    • Revocation supported via SPIRE
1
2
3
4

# Registration with 1-hour TTL
spire-server entry create \
  -spiffeID spiffe://example.com/agent/my-agent \
  -ttl 3600
  1. SPIRE Entry Deletion (CC6.2)
    • Remove SPIRE registration to revoke access permanently
1
2

# Revoke access
spire-server entry delete -spiffeID spiffe://example.com/agent/old-agent

CC7.2 - System Monitoring

Requirement: The entity monitors system components and the operation of those components for anomalies.

AgentWeave Controls:

  1. Audit Logging (CC7.2)
    • All requests logged with caller/callee identity
    • Authorization decisions logged
    • Failed access attempts logged
1
2
3
4

observability:
  audit_log:
    enabled: true
    destination: "syslog"
  1. Metrics (CC7.2)
    • Authorization success/failure rates
    • SVID rotation events
    • Capability invocation counts
  2. Alerting (CC7.2)
    • Alert on high denial rates
    • Alert on SVID rotation failures
    • Alert on unknown caller attempts

CC8.1 - Change Management

Requirement: The entity authorizes, designs, develops or acquires, configures, documents, tests, approves, and implements changes to infrastructure and software.

AgentWeave Controls:

  1. Configuration Validation (CC8.1) 
    1
2

    # Validate before deployment
agentweave validate config/production.yaml
  2. Version Control (CC8.1)
    • Policies stored in Git
    • Agent configurations in version control
    • Audit trail of changes
  3. Policy Testing (CC8.1) 
    1
2

    # Test policies before deployment
opa test policies/ -v

SOC 2 Evidence Collection

What to Collect:

  1. Configuration Files
    • Agent YAML configurations showing security settings
    • OPA policy files
    • SPIRE registration entries
  2. Audit Logs
    • Sample of authorization decisions
    • Access denied events
    • SVID rotation events
  3. Monitoring Dashboards
    • Screenshots showing security metrics
    • Alert configurations
    • Incident response logs
  4. Procedures
    • SVID rotation policy (document TTL settings)
    • Access revocation process
    • Policy review schedule

HIPAA (Health Insurance Portability and Accountability Act)

Overview

HIPAA requires protecting Protected Health Information (PHI) with administrative, physical, and technical safeguards.

Relevant Safeguards

164.312(a)(1) - Access Control

Requirement: Implement technical policies and procedures for electronic information systems that maintain ePHI to allow access only to those persons or software programs that have been granted access rights.

AgentWeave Controls:

  1. Unique User Identification (164.312(a)(2)(i))
    • Each agent has unique SPIFFE ID
    • Cannot share credentials (cryptographic identity)
  2. Automatic Logoff (164.312(a)(2)(iii))
    • SVIDs expire automatically (1 hour TTL)
    • Must re-authenticate after expiry
  3. Encryption and Decryption (164.312(a)(2)(iv))
    • All communication encrypted with TLS 1.3
    • Perfect forward secrecy

164.312(b) - Audit Controls

Requirement: Implement hardware, software, and/or procedural mechanisms that record and examine activity in information systems that contain or use ePHI.

AgentWeave Controls:

  1. Comprehensive Audit Logging
    • All PHI access logged
    • Who accessed what, when
    • Authorization decisions logged
1
2
3
4
5
6
7
8
9
10
11
12

observability:
  audit_log:
    enabled: true
    level: "info"
    fields:
      - "timestamp"
      - "caller_spiffe_id"
      - "callee_spiffe_id"
      - "capability"
      - "action"
      - "decision"
      - "trace_id"
  1. Log Retention
    • Logs sent to SIEM for retention
    • Configure retention per HIPAA requirements (6 years)

164.312(c) - Integrity Controls

Requirement: Implement policies and procedures to protect ePHI from improper alteration or destruction.

AgentWeave Controls:

  1. Data Integrity (164.312(c)(1))
    • TLS provides message integrity (MAC)
    • Tampering detected and rejected
  2. Digital Signatures (164.312(c)(2))
    • SVID is cryptographically signed
    • Cannot forge identity

164.312(d) - Person or Entity Authentication

Requirement: Implement procedures to verify that a person or entity seeking access to ePHI is the one claimed.

AgentWeave Controls:

  1. Strong Authentication
    • Mutual TLS authentication
    • Cryptographic proof of identity
    • Cannot impersonate without private key

164.312(e) - Transmission Security

Requirement: Implement technical security measures to guard against unauthorized access to ePHI that is being transmitted over an electronic communications network.

AgentWeave Controls:

  1. Encryption (164.312(e)(2)(i))
    • TLS 1.3 encryption for all transmissions
    • No plaintext communication possible
  2. Integrity Controls (164.312(e)(2)(ii))
    • Message authentication codes
    • Detect tampering in transit

HIPAA Implementation Recommendations

For PHI-Handling Agents:

  1. Enable payload redaction in logs: ```yaml observability: audit_log: include_payloads: false # Don't log PHI redact_fields:
    • "ssn"
    • "patient_id"
    • "diagnosis" ```
  2. Encrypt PHI at rest (application layer): 
    1
2
3
4
5
6
7

    class PHIProcessor(SecureAgent):
 @capability(name="process_phi")
 async def process(self, encrypted_phi: str):
     # Decrypt only in memory
     phi = self.decrypt(encrypted_phi)
     # Process...
     # Never log PHI
  3. Implement access controls: 
    1
2
3
4
5

    # Only allow healthcare providers to access PHI
allow if {
 input.caller_spiffe_id in data.healthcare_providers
 input.action in ["read_phi", "write_phi"]
}

PCI DSS (Payment Card Industry Data Security Standard)

Overview

PCI DSS protects cardholder data with requirements for network security, access control, and monitoring.

Relevant Requirements

Requirement 1: Install and Maintain Network Security Controls

AgentWeave Controls:

  1. Network Segmentation (1.2.1)
    • Use Kubernetes NetworkPolicies
    • Isolate agents handling card data
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: payment-agent-netpol
spec:
  podSelector:
    matchLabels:
      app: payment-processor
  policyTypes:
    - Ingress
    - Egress
  ingress:
    # Only from authenticated agents
    - from:
        - podSelector:
            matchLabels:
              pci-zone: "true"

Requirement 2: Apply Secure Configurations

AgentWeave Controls:

  1. Secure Defaults (2.2)
    • Security cannot be disabled
    • TLS 1.3 enforced
    • Default deny authorization

Requirement 4: Protect Cardholder Data with Strong Cryptography

AgentWeave Controls:

  1. Encryption in Transit (4.2)
    • TLS 1.3 for all communication
    • Strong cipher suites only
1
2
3
4
5

transport:
  tls_min_version: "1.3"
  cipher_suites:
    - "TLS_AES_256_GCM_SHA384"
    - "TLS_AES_128_GCM_SHA256"

Requirement 7: Restrict Access to System Components

AgentWeave Controls:

  1. Need-to-Know Access (7.1)
    • Default deny authorization
    • Grant minimal permissions
1
2
3
4
5
6

# Only payment gateway can call payment processor
allow if {
    input.caller_spiffe_id == "spiffe://example.com/agent/payment-gateway"
    input.callee_spiffe_id == "spiffe://example.com/agent/payment-processor"
    input.action in ["authorize", "capture"]
}

Requirement 8: Identify Users and Authenticate Access

AgentWeave Controls:

  1. Unique IDs (8.1)
    • Each agent has unique SPIFFE ID
  2. Strong Authentication (8.3)
    • Cryptographic authentication (mTLS)
    • Cannot share credentials

Requirement 10: Log and Monitor All Access

AgentWeave Controls:

  1. Audit Trail (10.2)
    • All access to cardholder data logged
    • Failed access attempts logged
    • Authorization decisions logged
  2. Log Details (10.3)
    • User identification (SPIFFE ID)
    • Type of event (capability called)
    • Date and time
    • Success or failure
    • Origination of event
1
2
3
4
5
6
7
8
9

{
  "timestamp": "2024-01-15T10:30:00Z",
  "caller_spiffe_id": "spiffe://example.com/agent/payment-gateway",
  "callee_spiffe_id": "spiffe://example.com/agent/payment-processor",
  "capability": "process_payment",
  "action": "authorize",
  "decision": "allow",
  "trace_id": "abc123"
}
  1. Log Review (10.6)
    • Send logs to SIEM
    • Set up automated analysis

GDPR (General Data Protection Regulation)

Overview

GDPR protects personal data of EU residents with requirements for consent, access, deletion, and security.

Relevant Articles

Article 32: Security of Processing

Requirement: Implement appropriate technical and organizational measures to ensure a level of security appropriate to the risk.

AgentWeave Controls:

  1. Encryption (Article 32(1)(a))
    • TLS 1.3 encryption in transit
    • Application responsible for encryption at rest
  2. Pseudonymization (Article 32(1)(a))
    • SPIFFE IDs (pseudonymous identifiers)
    • Can correlate via SPIRE but not from IDs alone
  3. Confidentiality (Article 32(1)(b))
    • Access control via OPA policies
    • Default deny
  4. Availability (Article 32(1)(b))
    • Circuit breakers prevent cascading failures
    • Rate limiting prevents DoS

Article 30: Records of Processing Activities

Requirement: Maintain records of processing activities.

AgentWeave Controls:

  1. Audit Logging
    • Record all data processing activities
    • Who processed what personal data
1
2
3
4

observability:
  audit_log:
    enabled: true
    # Log processing activities

Article 33: Notification of Personal Data Breach

Requirement: Notify supervisory authority of breach within 72 hours.

AgentWeave Controls:

  1. Breach Detection
    • Monitor for unauthorized access attempts
    • Alert on unusual patterns
1
2
3
4
5

# Alert on authorization failures
- alert: PossibleDataBreach
  expr: rate(agentweave_authz_denied_total[5m]) > 20
  annotations:
    summary: "Possible data breach - high access denial rate"
  1. Audit Trail for Investigation
    • Full logs of who accessed what
    • Trace IDs for incident investigation

GDPR Implementation Recommendations

For Personal Data Processing:

  1. Data Minimization (Article 5(1)(c)): 
    1
2
3
4

    @capability(name="get_user")
async def get_user(self, user_id: str, fields: list[str]):
 # Only return requested fields, not everything
 return await self.db.get_user(user_id, fields=fields)
  2. Right to Erasure (Article 17): 
    1
2
3
4
5

    @capability(name="delete_user")
async def delete_user(self, user_id: str):
 # Implement deletion capability
 await self.db.delete_user(user_id)
 self.logger.info(f"Deleted user {user_id}")
  3. Data Portability (Article 20): 
    1
2
3
4

    @capability(name="export_user_data")
async def export_user_data(self, user_id: str) -> dict:
 # Return data in structured format
 return await self.db.export_user(user_id)

FedRAMP (Federal Risk and Authorization Management Program)

Overview

FedRAMP standardizes security assessment for cloud services used by US federal agencies.

Relevant Controls (NIST SP 800-53)

AC-2: Account Management

AgentWeave Controls:

  • Unique accounts (SPIFFE IDs)
  • Automatic credential expiration (SVID TTL)
  • Revocation capability (delete SPIRE entries)

AC-3: Access Enforcement

AgentWeave Controls:

  • Default deny authorization
  • Policy-based access control (OPA)
  • Cannot bypass authorization

AC-17: Remote Access

AgentWeave Controls:

  • mTLS for all remote communication
  • Cryptographic authentication
  • Encryption in transit

AU-2: Audit Events

AgentWeave Controls:

  • Audit all authorization decisions
  • Log successful and failed access
  • Log administrative actions

AU-3: Content of Audit Records

AgentWeave Controls:

Audit logs include:

  • Event type (capability invoked)
  • Date and time
  • User identity (SPIFFE ID)
  • Outcome (allow/deny)
  • Source (caller SPIFFE ID)

IA-2: Identification and Authentication

AgentWeave Controls:

  • Unique identification (SPIFFE IDs)
  • Multifactor authentication (certificate + private key)
  • Cryptographic authentication (mTLS)

SC-8: Transmission Confidentiality and Integrity

AgentWeave Controls:

  • TLS 1.3 encryption
  • Message integrity (MAC)
  • No plaintext transmission

SC-12: Cryptographic Key Establishment

AgentWeave Controls:

  • Automated key management (SPIRE)
  • Regular rotation (SVID TTL)
  • No manual key distribution

FedRAMP Implementation

For FedRAMP authorization:

  1. Enable FIPS mode (if required): 
    1
2
3

    transport:
  fips_mode: true
  tls_min_version: "1.2"  # FIPS 140-2 validated
  2. Continuous monitoring: 
    1
2
3
4
5
6
7
8

    observability:
  metrics:
 enabled: true
  audit_log:
 enabled: true
 destination: "syslog"
  tracing:
 enabled: true
  3. Incident response:
    • Documented procedures for security events
    • Integration with SIEM
    • Alerting on security events

Compliance Checklist

Use this checklist to assess compliance readiness:

Authentication & Identity

  • SPIRE deployed and configured
  • Unique SPIFFE ID per agent
  • SVID TTL ≤ 1 hour
  • Platform attestation enabled
  • Credential rotation automated

Authorization

  • Default deny policy enforced
  • OPA policies documented
  • Least privilege applied
  • Policy testing in place
  • Policy review schedule defined

Encryption

  • TLS 1.3 enforced
  • mTLS for all communication
  • Strong cipher suites only
  • Application-layer encryption for sensitive data at rest

Audit Logging

  • Audit logging enabled
  • Logs include required fields
  • Logs sent to SIEM
  • Log retention configured
  • Log review process defined

Monitoring

  • Security metrics collected
  • Alerts configured
  • Dashboards created
  • Anomaly detection enabled

Incident Response

  • Incident response plan documented
  • Revocation process defined
  • Investigation procedures defined
  • Communication plan defined

Documentation

  • Security controls documented
  • Configuration documented
  • Policies documented
  • Procedures documented
  • Evidence collected

Evidence for Auditors

When preparing for compliance audits, collect:

Technical Evidence

  1. Configuration Files
    • Agent configurations showing security settings
    • OPA policy files
    • SPIRE registration entries
  2. Logs
    • Sample audit logs
    • Authorization decisions
    • SVID rotation events
  3. Monitoring Screenshots
    • Security dashboards
    • Alert configurations
    • Metric graphs

Process Evidence

  1. Procedures
    • Access provisioning process
    • Access revocation process
    • Policy review process
    • Incident response plan
  2. Training Records
    • Security training for operators
    • Policy development training
  3. Review Records
    • Policy review logs
    • Access review logs
    • Audit log review records

Important Notes

  1. AgentWeave provides controls, not compliance:
    • You must implement organizational policies
    • You must conduct regular audits
    • You must train personnel
  2. Application-layer security required:
    • AgentWeave secures agent-to-agent communication
    • Your application must protect data at rest
    • Your application must implement business logic security
  3. Consult compliance experts:
    • Each organization's compliance needs are unique
    • Regulations are subject to interpretation
    • Work with qualified compliance professionals

Next Steps

References