Build content, webapps, and documentation based on the selfdriven Digital Interaction Framework — the layered trust architecture for digital relationships. Use when the user mentions trust layers, digital interaction framework, relationship lifecycle, agent network, trust tiers, policy stacks, risk calibration, transactional vs interactional, fidelity/confidence/provenance, or Architecture Decision Records (ADRs) in the selfdriven context.
Build interactive info sites, documentation, diagrams, and reference materials based on the selfdriven Digital Interaction Framework — a layered trust architecture for self-sovereign digital relationships.
The Digital Interaction Framework defines how verifiable digital relationships are established, maintained, and dissolved using KERI/ACDC cryptographic infrastructure. It organises trust into five interdependent layers, a six-stage relationship lifecycle, a tiered agent network, a governance model with Architecture Decision Records, and four interconnected policy domains.
This skill should be used alongside the selfdriven-ecosystem skill which provides the brand system, design tokens, typography, and visual language. This skill provides the conceptual framework and content model.
Consult this skill for any mention of:
Every digital interaction rests on a stack of interdependent layers. Each layer builds assurance and reduces risk for the layers above. Layers are numbered bottom-up (01 = foundation, 05 = surface).
The most stable layer, changing infrequently. Four interconnected policy domains:
| Domain | Scope | Key Elements |
|---|---|---|
| Technical | Infrastructure & security | Ed25519 signing, X25519 exchange, AES-256-GCM at rest, TLS 1.3 in transit, HSM key storage, 2-of-3 witness topology, OpenAPI 3.1 contracts, 7yr audit retention |
| Financial | Economic & regulatory | Transaction limits, velocity controls, NPP/BPAY/SWIFT settlement, AUSTRAC AML/CTF, APRA CPS 234, CDR consent, cross-border governance |
| Human | Conduct & privacy | Conductor 70/30 model, role-to-credential access mapping, graduated disclosure, OAIC compliance, escalation authority chains, consent management |
| Agent | Delegation & autonomy | Action whitelists, 24h default time limits, all ixn logged, anomaly/threshold escalation, ACDC-gated access, A2A protocol rules, L1–L5 classification |
Policy Domain Interactions — strong dependencies exist between: Technical ↔ Agent (infra constrains agent capability), Financial ↔ Human (limits constrain decisions), Human ↔ Agent (conductors set agent boundaries), Technical ↔ Financial (protocols enable compliance).
Rules, structures, and oversight that shape how technology is deployed.
dip eventsdip — scope boundaries, time limits, escalation triggers, human-in-the-loop checkpointsThe cryptographic substrate. Eight technology domains, each with specific detail:
| Domain | Detail |
|---|---|
| AuthN / AuthZ | FIDO2/WebAuthn passkeys bound to KERI AIDs via ixn. ACDC-derived authorisation scopes. Passkey public key anchored to AID via ixn event. |
| Protocols | KERI events (icp, rot, ixn, dip, drt), OOBI discovery, A2A agent-to-agent communication |
| Encryption | AES-256-GCM at rest, TLS 1.3 in transit, X25519 end-to-end agent messages |
| Cryptology | Ed25519 signing, X25519 key exchange, SHA-256 event digests, SHA-512/HKDF key derivation |
| Privacy | ACDC graduated disclosure, selective correlation control, protocol-level data minimisation |
| Path History | Complete identifier traversal record, ixn sequence gap detection, forensic KEL replay |
| Key Management | Mandatory pre-rotation, HSM integration for high-value AIDs, M-of-N multi-sig thresholds |
| Witness Network | 2-of-3 default (high-value: 3-of-5), geographic distribution across 2+ AZs, watcher duplicity detection |
Establishing the level of risk tolerance for each interaction:
The interaction itself — where value is exchanged:
| Layer | Rate of Change | Primary Actor | KERI Primitive | Failure Mode |
|---|---|---|---|---|
| 05 — Interaction | Per-exchange | Both parties | ixn | Context mismatch, scope creep |
| 04 — Trust | Progressive | Verifier | ACDC chain | Fidelity degradation, confidence erosion |
| 03 — Technology | On rotation | Infrastructure | icp, rot | Key compromise, witness failure |
| 02 — Governance | Periodic | Human Conductor | dip, rules | Authority gap, escalation failure |
| 01 — Policy | Infrequent | Foundation | ADR | Regulatory non-compliance |
Six stages from discovery to dissolution. Each stage maps to specific KERI primitives and trust requirements.
icp → AID genesis, ACDC → capability publish, oobi → endpoint discoveryixn → context binding, ACDC → credential present, TEL → issuance verifydip events establish delegated AIDs with scoped authoritydip → agent delegation, ACDC → terms credential, ixn → agreement anchorixn → operation log, TEL → status check, ACDC → re-presentrot events; pre-rotation means next keys already committedrot → key rotation, ACDC → credential reissue, ixn → terms updatedrt; agent authority terminatedTEL → revoke, drt → delegation revoke, KEL → permanent record| Stage | Trust Level | Credential Flow | Governance | Reversibility |
|---|---|---|---|---|
| Discovering | None | Outbound publish | Minimal | Full |
| Co-Creating | Emerging | Bilateral exchange | Light | High |
| Proposing | Threshold | Terms codification | Moderate | Moderate |
| Using | Established | Continuous verify | Active | Low |
| Updating | Maintained | Refresh & reissue | Periodic | Moderate |
| Archiving | Residual | Revocation | Wind-down | Irreversible |
Tiered identity architecture connecting individual AIDs to organisational networks.
The self-certifying root identifier. All trust relationships radiate from this node.
Highest trust. Full bilateral credential chain verification, mutual KEL witnessing, direct signed communication, real-time TEL status, eligible for agent delegation.
Trust derived from group membership via shared issuer. Group credential schema membership, issuer-mediated trust, attribute-based access control, progressive elevation to Tier 1.
The core operational unit of the Human Conductor model. Conductor holds root AID; agent operates under delegated AID via dip. 70/30 energy split. All agent actions logged via ixn.
Always-on network infrastructure. Witness nodes (2-of-3), watcher nodes (KEL monitoring/duplicity detection), discovery endpoints (OOBI resolution), TEL registries (credential status broadcasting). Geographic distribution required.
| Protocol | Purpose |
|---|---|
| KEL | Append-only key event log. Events: icp, rot, ixn, dip, drt |
| TEL | Transaction Event Log — credential lifecycle tracking (issuance, revocation, status) |
| OOBI | Out-of-Band Introduction — discovery protocol bootstrapping trust via endpoint resolution |
Seven foundational ADRs govern the framework. Each follows Context → Decision → Consequences structure.
| ADR | Title | Scope | Key Decision |
|---|---|---|---|
| ADR-001 | KERI-First Identity | All domains | All identifiers are KERI AIDs. No alternative primary identity scheme. External identifiers (email, DID:web) as secondary bindings only. |
| ADR-002 | Passkey-Only AuthN | All user-facing apps | FIDO2/WebAuthn passkeys bound to KERI AIDs. No passwords, no SMS OTP, no email magic links. |
| ADR-003 | ACDC Credential Schema | All credentials | Foundation-governed ACDC schemas. Graduated disclosure by default. Schema changes require governance review. |
| ADR-004 | Witness Threshold 2/3 | Infrastructure | 3 witnesses, 2-of-3 confirmation. High-value: 5 witnesses, 3-of-5. Geographic distribution across 2+ AZs. |
| ADR-005 | Agent Delegation via dip | Agent operations | All agent authority via KERI dip events with explicit scope, time window, interaction limits, and escalation triggers. |
| ADR-006 | Pre-Rotation Mandatory | All AIDs | Next rotation keys committed at inception. Not optional. Enables instant recovery without authority dependency. |
| ADR-007 | TEL for All Credentials | All ACDC credentials | Every ACDC has a TEL entry. Verifiers check TEL before accepting any presentation. |
Five-level classification for AI agent independence:
| Level | Name | Authority | Human Oversight | Example |
|---|---|---|---|---|
| L1 | Assistive | Read-only, suggest | Continuous | Data retrieval, report generation |
| L2 | Supervised | Execute with approval | Per-action | Draft communications, form filling |
| L3 | Bounded | Execute within rules | Exception-based | Routine transactions, credential issuance |
| L4 | Autonomous | Self-directed in scope | Periodic review | Service monitoring, anomaly response |
| L5 | Strategic | Cross-domain coordination | Outcome-based | Multi-agent orchestration, resource allocation |
When building interactive framework documentation sites:
Use an HTML <canvas> element with requestAnimationFrame for the agent network visualisation:
#2d8a4e, radius ~14px, label "R"#4a7fb5, radius ~10px, label "R"#9e9085, radius ~7px, no labelrgba(200,68,47,0.12)) with animated packets travelling along themMath.sin(time * 1.5 + i) * 3IntersectionObserver (only start animation when visible)JetBrains Monojavascript:void(0) with data-target attributes and scrollIntoView({behavior:'smooth'}) for nav links — avoids popup behaviour in artifact viewersclassList.toggle('active-link')<audio> element with preload="none"https://selfdriven.foundation, .money, .pro, .network, .ai, .bot, .insure with target="_blank"Before delivering Digital Interaction Framework content:
icp, rot, ixn, dip, drt, not generic termsselfdriven-ecosystem skill applied (Poppins, JetBrains Mono, flamingo accent, dark mode)