Home

Integrity Protocol Wiki

Compiled knowledge base for the Integrity Protocol monorepo — a trust/ compliance layer for AI agents on Base L2. This page is the map; every section below links to the real page. Governance/conventions: WIKI_SCHEMA.md (page format), this index.md (the canonical catalog), WIKI_INDEX.md (a retained legacy catalog), WIKI_LOG.md (chronological history, append-only). Cross-package decisions live in ../INTERFACE_CONTRACT.md; how this wiki gets kept in sync with the code is ../../AGENTS.md.

Start here if you're new: The Four Foundational Primitives (the concepts), then Agent Primitives (the 7 per-agent contracts every other page assumes you understand) and Persistent Memory (the continuity primitive that gates registration), then AIS (the trust score) and Telemetry Ingestion Pipeline (how agent behavior becomes that score).

Specification authority — 2026-09-05 (corrected; see docs/DOCUMENT_STATUS.yaml)

• docs/SPEC.md (v1.0.0-draft) is the accepted normative specification. docs/DOCUMENT_STATUS.yaml is the single machine-readable source of truth for this — check it directly rather than trusting this page to stay current.
• Integrity Protocol v0.4 is now archived, not normative. A previous version of this section (dated 2026-08-29) still named it as the accepted baseline after the real cutover to docs/SPEC.md had already happened for at least Phase III adapter-registry work — see PRODUCTION_GAPS.md's dated pointer-correction entry for the drift this closes.
• Integrity Protocol v0.5 proposed is likewise archived, not a live review candidate.
• Whitepaper v3.2 is the current explanatory and non-normative whitepaper.
• The v3.2 PDF is generated publication output from the archived v3.2 source; code, tests, and deployments are separate implementation evidence layers.

Current cross-repository audit — 2026-08-25

Note, 2026-08-12: the ecosystem is a three-repository model now — integrity-mvp (linked below as a dated historical pointer) has been superseded by integrity-dashboard/ inside this repository; see Ecosystem Dependencies for the current correction. The list below is left as-recorded from the 2026-08-06 audit, not rewritten.

The cross-repository implementation plan and audit ledger are maintained in Repository Implementation Plans. Repository-local status pages are:

• integrity-core audit status: historical strong-testnet-prototype assessment; production readiness was not established.
• integrity-mvp audit status: frontend build and unit tests pass; lint and dependency-security gaps remain.
• xibalba-shield audit status: Linux-first prototype; 2 of 3 eBPF probes verified; TCP-connect blocked.
• xibalba-cortex audit status: local MCP memory prototype; tests pass with Drive extras; active worktree changes require review.

These pages distinguish DONE, PARTIAL, PLANNED, BLOCKED, UNVERIFIED, and REQUIRES REVIEW. Historical wiki log entries and design records remain historical evidence.

The latest compiled session evidence keeps Phase I contract work and Hermes/BCC live dispatch in PARTIAL/BLOCKED status; local adapter validation and policy-service reachability do not establish live provider parity or host-tool execution.

System at a glance

flowchart TB
    Wallet["Agent's own wallet"]
    Factory["AgentPrimitivesFactory"]

    subgraph OnChain["On-chain (EVM / Base Sepolia + anvil)"]
        SA["SovereignAgent<br/>(identity account)"]
        StA["StateAnchor<br/>(per-agent audit root)"]
        subgraph Clones["5 EIP-1167 minimal-proxy clones"]
            RR["ReputationRegistry"]
            SL["Slasher"]
            VR["VerifierRegistry"]
            CG["ComplianceGate"]
            AP["AgentProfile"]
        end
    end

    SDK["integrity-sdk / integrity-cli"]
    BCC["bcc_middleware<br/>(FastAPI + OPA)"]
    Oracle["integrity-oracle<br/>(Rust/Axum)"]
    Dashboard["integrity-dashboard<br/>(React + Python)"]
    UserAPI["integrity-userapi<br/>(FastAPI + Postgres)"]

    Wallet -->|signs direct deploys| SA
    Wallet -->|signs direct deploys| StA
    Factory -->|clones| RR
    Factory -->|clones| SL
    Factory -->|clones| VR
    Factory -->|clones| CG
    Factory -->|clones| AP

    SDK --> Wallet
    SDK -->|pre-execution gate| BCC
    SDK -->|signed telemetry| Oracle
    BCC -->|reads AIS, pushes score on-chain| Oracle
    Oracle -->|resolve + score| OnChain
    Oracle --> Dashboard
    UserAPI -->|read-only fan-out, never chain| Oracle
    Dashboard --> UserAPI

Loading

Table of contents

Concepts — identity & on-chain primitives

• The Four Foundational Primitives — the concepts the protocol rests on; start here
• Agent Primitives (Self-Sovereign Identity) — the 7 per-agent contracts (a different sense of the word)
• Persistent Memory, Genesis Root & Lineage — foundational primitive: no agent registers without an anchored genesis memory root ([PARTIALLY BUILT])
• Decentralized Identifier (DID)
• Identity Ceiling & Verification Ladder — [PARTIALLY BUILT]

Concepts — trust & scoring

• Agent Integrity Score (AIS) — the formula + oracle's server-side re-derivation trust model
• Telemetry Ingestion Pipeline — end-to-end: SDK collection → batching → signing → the oracle's 11-step ordered pipeline → AIS
• Local Metrology — the exact entropy/grounding/sacrifice/compliance formulas
• Observability & PHI Safety Pipeline — the Redactor, redact_phi, the oracle-side PHI backstop

Concepts — behavioral gating & cryptography

• Behavioral Commitment Chain (BCC) — the pre-execution signed-intent gate
• Invocation Correlation Profile v1 — canonical per-attempt UUID joining intent, policy, execution, and effect without content-hash ambiguity
• Merkle Batching & Anchoring Convention
• Zero-Knowledge Proving Pipeline (ZKP)

Concepts — compliance & markets

• ComplianceGate & Integrity Health — the HIPAA/healthcare vertical
• Smart BAA — on-chain Business Associate Agreement escrow
• Integrity Market — prediction markets, binary options, A2A capital allocation

Concepts — wire spec & testing

• AIS API — Versioned Wire Spec — the generated, externally-supported /v1/* spec
• Testing Strategy — the 3-layer test pyramid

Concepts — planned / design-only

• Identity Ceiling & Verification Ladder — [PARTIALLY BUILT]
• Cross-Chain Reputation Sync — [PLANNED]
• A2A Negotiation Protocol — [PLANNED]
• ZK-ML Model-Inference Verification — [PLANNED]

Entities — one page per real package

• contracts — Solidity/Foundry: the 7 primitives, factory, registries, XNS, Integrity Health, market layer, $ITK
• integrity-oracle — Rust/Axum: AIS scoring, server-side telemetry re-derivation, on-chain reads, markets/leaderboard
• integrity-sdk — Python agent library: identity, BCC, markets, telemetry, PHI redaction
• integrity-cli — developer CLI, independent reimplementation of the SDK's core flows
• bcc_middleware — FastAPI + OPA pre-execution policy gate + reputation-sync loop
• integrity-userapi — FastAPI + Postgres user accounts/auth, strictly non-chain
• integrity-dashboard — the React/Vite dashboard + demo/ scenario engine
• integrity-zkp — the real Noir/Barretenberg circuit

Architecture

• Ecosystem Dependencies — cross-repository ownership and dependency direction for integrity-core (including its integrity-dashboard/ component), Xibalba Cortex, and Xibalba Shield.
• Repository Implementation Plans — closed/planned/blocked implementation ledger for integrity-core, integrity-dashboard/, Xibalba Shield, and Xibalba Cortex.

Guides

• Smart Contract Development — writing/testing/deploying a new contract
• Multi-Domain Guardrails Design — [DESIGN, PARTIALLY BUILT]

Reference

• WIKI_INDEX.md — full catalog, one-line description per page (the canonical index)
• WIKI_LOG.md — chronological record of every wiki change, append-only
• WIKI_SCHEMA.md — page format, frontmatter, tag taxonomy

Open queries

• No LLM-as-judge rubric exists anywhere in this repo — the judge_evaluations ingestion schema is built but the scoring rubric is an open product question. See Observability & PHI Safety.

Acronym glossary

• AIS — Agent Integrity Score → concepts/ais.md
• BAA — Business Associate Agreement → concepts/smart-baa.md
• BCC — Behavioral Commitment Chain → concepts/bcc.md
• DID — Decentralized Identifier → concepts/did.md
• VTL — (old term) Verifiable Trust Layer → see Observability & PHI Safety for what's actually built
• ZKP — Zero-Knowledge Proof(ing pipeline) → concepts/zkp.md

No aspirational content

Every page here documents what exists in the code right now. A feature described in a spec but not yet implemented is explicitly marked [PLANNED] or [DESIGN, PARTIALLY BUILT] in its title/index entry — never written as if it's real. See WIKI_SCHEMA.md for the full convention.