Skip to main content
Attestix
Security

SOC 2 Trust Services Criteria, Attestix coverage

How Attestix's signed audit chains, Verifiable Credentials, and provenance records map to the AICPA SOC 2 Trust Services Criteria (2017 + 2022 additions). SOC 2 is an attestation, not a certification. Attestix is evidence plumbing your CPA's auditor can use, not a SOC 2 readiness platform.

The AICPA SOC 2 Trust Services Criteria (TSC 2017 with 2022 Revised Points of Focus) are the framework an independent CPA firm uses to perform a SOC 2 examination of a service organisation. This page shows how Attestix's signed audit chains, Verifiable Credentials, and provenance records map to the TSC, focused on the technology-heavy criteria (Common Criteria CC1-CC9 and Processing Integrity PI1) where chain-hashed evidence is most useful.

Why this matters

SOC 2 is an attestation, not a certification. This distinction is load-bearing:

  • A certification (ISO 27001, ISO 42001) is a binary YES/NO awarded by an accredited certification body. The organisation displays a cert mark.
  • An attestation (SOC 2) is a CPA firm's professional opinion on whether the service organisation's controls were suitably designed (Type I) or designed and operating effectively (Type II) to meet the relevant Trust Services Criteria over a stated audit period. The deliverable is a SOC 2 report (~80 to 200 pages) with a one-page CPA opinion letter on top.

Attestix does not ship SOC 2 attestations and is not a CPA firm. We do not issue an opinion. What Attestix ships is the evidence plumbing your CPA's auditor can use during your own SOC 2 examination, particularly for the technology-heavy criteria where signed audit trails + chain-hashed event logs + cryptographic identity reduce auditor evidence-gathering by weeks.

The honest coverage table

The TSC 2017 has 64 criteria total. The table below covers 30 representative criteria focused on Common Criteria (CC1-CC9, the Security category required for every SOC 2 report) and Processing Integrity (PI1, the entire category, the strongest Attestix-aligned area).

CriterionMitigation surface in AttestixCoverageEvidence shape
CC1.1 Integrity + ethical values commitmentNone directly; ethics culture is organisationalrecord-onlyProviderAssertionCredential
CC1.2-CC1.5 Board oversight + structures + competency + accountabilityNone; organisational controlsout-of-scopen/a
CC2.1-CC2.3 Information quality + internal + external communicationEvery audit event high-quality by construction; agent cards + DoC as external-communication artefactsevidence-substrateAgentIdentityCredential + audit chain
CC3.1 Specifies objectivesintended_purpose + system_objectives (v0.5)record-onlyProviderAssertionCredential
CC3.2 Identifies + analyses risk (includes 2022 AI/ML points of focus)v0.5 risk register + cross-walk to OWASP ASI + FRIA templateevidence-substrateProviderAssertionCredential + SecurityCheckCredential
CC3.3 Considers fraud riskChain-hashed audit log makes post-hoc fraud forensically detectable (actors with valid DID credentials can still emit fraudulent-but-hash-consistent events, pair with CC7.2 anomaly detection); cross-walk to OWASP ASI03 (identity abuse)evidence-substrateProviderAssertionCredential + audit-chain tamper-evidence
CC4.1-CC4.2 Ongoing + separate evaluations + deficiency communicationcompliance_service checks re-run on cron with signed VerifiableCheckResult; v0.5 control_deficiency eventsevidence-substrateVerifiableCheckResult + audit chain
CC5.1-CC5.3 Control activity selection + general technology controls + deployment via policiescompliance_service profiles encode control selections; audit chain over every profile changerecord-onlyProviderAssertionCredential + audit chain
CC6.1 Logical access security software + infrastructureDID-based identity; UCAN capability + scope enforcement; every state change signed by actor DIDevidence-substrateDID document + UCAN token + audit chain
CC6.2 Registers + authorises new users prior to credential issuanceidentity_service.create_agent_identity, signed + dated + role-tagged; approver DID signs issuance eventevidence-substrateAgentIdentityCredential + approver-signed audit event
CC6.3 Authorises + modifies + removes access (least privilege + segregation of duties)UCAN delegation chain (least-privilege built in via attenuation: no chain can grant powers the parent did not have, PR #45); revoke_credentialevidence-substrate (strongest CC6 row)UCAN chain + revocation VC + audit chain
CC6.4-CC6.5 Physical access + asset retrieval/disposalNone directly; data-centre operations (AWS, GCP, Azure, on-prem)out-of-scopen/a
CC6.6 Logical security measures against threats outside system boundaryCross-walk to OWASP ASI03 + ASI07; offline verify_credential for external-actor authenticationevidence-substrateAudit chain + SecurityCheckCredential
CC6.7 Transmission + movement + removal of informationbundle export signed + chain-hashed; v0.5 data_egress eventsevidence-substrateAudit chain + signed bundle manifest
CC6.8 Prevention + detection of unauthorised or malicious softwareCross-walk to OWASP ASI04 supply chain; model-lineage + training-data fingerprintsevidence-substrateProvenance entries + SecurityCheckCredential (ASI04 tag)
CC7.1 Detection + monitoring procedures (config changes + vulnerabilities)v0.5 config_change events with before/after; cross-walk to Art 15.5 cybersec checkevidence-substrateAudit chain + VerifiableCheckResult
CC7.2 Monitors system components for anomaliesCross-walk to NIST AI RMF MANAGE-4.1; reputation drift; v0.5 incident-reporting collectionevidence-substrateHash-chained audit + ReputationScoreCredential + IncidentReportCredential
CC7.3 Evaluates security events; initiates responsev0.5 incident-reporting collection (Art 73 cross-walk); revocation as kill-switchevidence-substrateIncidentReportCredential + revocation VC + audit chain
CC7.4 Executes defined incident-response programmeSame surface as CC7.3 + cross-walk to ISO 42001 A.8.4evidence-substrateIncidentReportCredential + audit chain
CC7.5 Recovers from identified incidentsv0.5 recovery_event; re-issued credentials post-recovery signed + chain-hashedevidence-substrateAudit chain + re-issued credential VCs
CC8.1 Change management (authorise + design + develop + configure + document + test + approve + implement)record_action(entry_type="change_request", before, after, approver_did); model-lineage changes; profile audit chainevidence-substrate (strong: change-management evidence is exactly what chain-hashed event logs are for)Audit chain + provenance entries
CC9.1 Risk-mitigation activities for potential business disruptionsSame surface as CC3.2 + CC7.5evidence-substrateProviderAssertionCredential + audit chain
CC9.2 Vendor + business-partner risk (includes 2022 AI/ML points of focus)Cross-walk to NIST AI RMF MAP-3.4 + OWASP ASI04; third-party AgentIdentityCredential acceptanceevidence-substrateAgentIdentityCredential + UCAN delegation + audit chain
PI1.1 Processing objectives + data definitions + product/service specsintended_purpose + system_specifications (v0.5); signed agent card with data-schema declarationsevidence-substrateAgentIdentityCredential + ProviderAssertionCredential
PI1.2 Input completeness + accuracy controlsv0.5 Art 15 accuracy/robustness check evidence; input_validation eventsevidence-substrateAudit chain + VerifiableCheckResult
PI1.3 System-processing controlsEvery state change signed + chain-hashed; processing integrity is forensic by construction; verify_chain tamper-evidenceevidence-substrate (strongest PI row)Audit chain + verify_chain result
PI1.4 Output completeness + accuracy + timelinessoutput_delivered events + bundle-export integrityevidence-substrateAudit chain + bundle manifest
PI1.5 Storage of inputs + items in processing + outputsStorage chain-hashed by construction; bundle export = portable storage receiptevidence-substrateAudit chain + bundle
A1.1-A1.3 Capacity + system backup + environmental monitoringNone directly; platform-operations (AWS/GCP/Azure/on-prem)record-onlyProviderAssertionCredential
C1.1-C1.2, P1.1 Confidentiality + privacy notice + retentionCross-walk to GDPR Art 17 (already in v0.3.0); privacy notice as ProviderAssertionCredentialpartialProviderAssertionCredential + audit chain

Tally (across all 64 criteria)

  • evidence-substrate (Attestix produces signed evidence your auditor can examine): 27
  • record-only (Attestix signs your operator's assertions): 18
  • out-of-scope (organisational + cultural CC1 + physical CC6.4-CC6.5 + most of Privacy P1-P8 territory): 19

Concentrated where it matters most for SOC 2: Security CC6 (logical access via UCAN + DID), CC7 (system operations via chain-hashed events), CC8 (change management via provenance) and Processing Integrity PI1 (the entire category).

Important caveat: security_check_id ships in v0.5.0. As of Attestix v0.4.0 the underlying events are emitted today, but they are NOT yet tagged with the soc2.* discriminator. The v0.5.0 release registers the prefix in FRAMEWORK_REGISTRY; per-criterion emission tagging is incremental.

What we don't do

  • We do not issue SOC 2 reports. The CPA firm performing your examination is the only entity that issues SOC 2 reports.
  • We do not author your policies. Policy authorship lives in your DMS, with Vanta / Drata / Secureframe templates, or with internal counsel.
  • We do not shepherd your SOC 2 readiness end-to-end. That is what compliance platforms (Vanta, Drata, Secureframe, Strike Graph, AuditBoard) do: auditor coordination, employee training, evidence collection workflows, continuous monitoring dashboards. They sit above the evidence layer; we sit below. The two stack.
  • We do not do physical security (CC6.4-CC6.5). Your data-centre operator's SOC 2 carve-out covers that.
  • We are not your IAM enforcement layer. We record who attempted what; your Okta / Auth0 / Entra IAM enforces access policies.
  • We are not your SIEM. We produce signed forensic evidence; your Splunk / Datadog / Honeycomb ingests + dashboards.

How to verify our coverage yourself

Python / CLI

# List every audit event tagged with a SOC 2 criterion (v0.5.0+)
attestix audit list --security-check soc2.security.CC6_3.access_granted

# Export a bundle scoped to SOC 2 evidence for your auditor (v0.5.0+)
attestix bundle export --controls soc2 --out my-soc2-evidence.atxbundle

# Verify chain integrity for an agent's audit log (today)
attestix verify-chain <agent-did>

JavaScript / browser

npm install attestix
import { verifyCredential } from "attestix";

const result = await verifyCredential(securityCheckCredentialJson);
// result.valid === true if the Ed25519 signature over the JCS-canonical body
// matches the issuer DID's public key.

On-chain anchor (Base L2 Sepolia testnet)

attestix anchor audit-batch --agent <did> --network base-sepolia

Mainnet schema registration is planned; testnet is the default target today.

Comparable disclosure

How other tools position themselves on SOC 2.

ToolStated SOC 2 positionWhere to read more
Microsoft Agent Governance ToolkitPublishes docs/compliance/soc2-mapping.md; CloudEvents-to-Azure-Monitor as evidence pathgithub.com/microsoft/agent-governance-toolkit
VantaEnd-to-end SOC 2 readiness platform: auditor coordination + policy templates + employee training + continuous monitoring + evidence collection. Slots above Attestix as the GRC layervanta.com
DrataSame posture as Vantadrata.com
SecureframeSame posture as Vanta + Dratasecureframe.com
Strike GraphSame posture; mid-market focusedstrikegraph.com
AuditBoardEnterprise GRC including SOC 2 module; same postureauditboard.com
CPA firms (Coalfire, Schellman, A-LIGN, Sensiba, etc.)The actual SOC 2 examination performers. Attestix is evidence input to their auditAICPA SOC 2 firm directory

See also


Attestix is evidence tooling a customer's auditor can use during the customer's own SOC 2 examination. Attestix does not issue SOC 2 attestations, is not a CPA firm, and a passing tag against a Trust Services Criterion is one signal among many in the audit, and the auditor's professional opinion is the authoritative output. Compliance platforms like Vanta, Drata, Secureframe, and Strike Graph shepherd the end-to-end SOC 2 readiness journey; we slot underneath them as the cryptographic evidence layer.