TessIndex ties autonomous agents and primitives to persistent, capability-verified identities—anchoring attestations on-chain while managing rich metadata off-chain—to enable accountable, discoverable and tokenizable agent commerce. The paper is a detailed architecture proposal but offers no implementation results or empirical validation.

Main Finding

TessIndex proposes a unified, capability‑verified identity system for the emerging agent economy that (1) provides persistent identities for not just agents but also tools, skills, services and other primitives, (2) replaces self-declared capability claims with predicate‑based cryptographic verification and execution evidence, and (3) links performance, creator identity, and commercial value (including tokenization) via a dual‑plane architecture: compact, immutable commitments on blockchain and rich, dynamic metadata in a centralized registry.

Key Points

• Three core gaps in current infrastructure:
  • Lack of systemic accountability (no persistent identity chain from human intent → agents → invoked primitives).
  • Unverified/self‑declared capability claims without execution evidence.
  • Missing linkage between creator identity, agent performance, and economic value.
• Dual‑plane architecture:
  • Blockchain plane: stores compact cryptographic commitments, ownership records, verification receipts, and anchors of trust (immutable proofs).
  • Centralized registry plane: stores rich, dynamic metadata (manifests, endpoints, performance logs, commerce metadata) for efficient discovery and integration.
• Six modular domains that compose the system:
• Identity — canonical identifiers, ownership binding, schema for primitives.
• Runtime — deterministic lifecycle, registration, resolution, revocation.
• Governance — record authority, automated publishing pipelines, lifecycle management.
• Interoperability — adapters to A2A, MCP, ACP, cross‑registry identity mapping and federated trust.
• Economy — intent‑based discovery, wallet binding, verify‑then‑pay flows, tokenization/assetization of agents.
• Trust & Security — predicate‑based capability verification, cryptographic task commitments, TEEs/trusted validators, abuse resistance and reputation.
• Capability verification model:
  • Predicate‑based verification replaces free‑text claims with verifiable predicates.
  • Verification receipts (cryptographic proofs) are anchored on‑chain; task execution evidence and verification history are recorded in the registry.
  • Supports proof of capability under declared operating conditions and ongoing execution records.
• Economic features:
  • Immutable linking of performance and reputation to persistent identities enables valuation and assetization (token issuance, trading, governance).
  • Supports escrowed payments and automated settlement tied to verified tasks.
• Interoperability & federation:
  • Designed to interoperate with existing registries and standards (e.g., NANDA, HOL Registry, ERC proposals) via adapters and identity mapping.
• Tradeoffs:
  • Balances on‑chain security and immutability with off‑chain scalability, expressiveness, and privacy via the centralized plane.
  • Potential centralization risks in the registry plane; governance mechanisms and federated mappings are proposed to mitigate this.

Data & Methods

• Nature of the contribution: system design / architecture paper (conceptual and protocol specification), not an empirical or experimental study.
• Methods and components described:
  • Architectural design of a two‑plane system (blockchain commitments + centralized registry).
  • Definition of object models, schemas, lifecycle semantics, and authority/authentication models.
  • Specification of a predicate‑based capability verification process, cryptographic receipts, and verification anchoring on the blockchain.
  • Integration patterns for interoperability (protocol adapters for agent communication standards like A2A, MCP, ACP).
  • Economic constructs: wallet binding, verify‑then‑pay flows, escrow, tokenization lifecycle, and reputation systems (on‑chain anchors + off‑chain reputation).
  • Security mechanisms: cryptographic commitments, trusted execution environment use cases, and access controls.
• No datasets, experiments, or quantitative evaluations are reported; validation is conceptual and architectural (design rationale, gap analysis, and specification).

Implications for AI Economics

• Reducing information asymmetry and moral hazard:
  • Verifiable capability claims and execution evidence lower uncertainty about agent performance; this should reduce adverse selection and allow more complex, higher‑value agent‑mediated transactions.
• Enabling valuation and liquidity of agent assets:
  • Persistent identities + immutable performance records enable agents (and agent‑linked projects) to be tokenized and valued like digital assets. This can create new markets (buy/sell/trade agent assets, derivatives, reputation markets).
• Lowering transaction and verification costs:
  • Standardized verification receipts and an identity root of trust reduce due diligence costs for delegating tasks to agents and for multi‑party settlements.
• Design space for incentives and market design:
  • Market designers must study incentive‑compatible verification (who verifies, how validators are rewarded, how to prevent gaming), reputation dynamics, and governance mechanisms to avoid centralization or capture.
• Network effects and market concentration:
  • A unified identity layer can create strong network effects (discovery, reputation). Without careful governance, this may lead to dominant registry operators or validator cartels—economic concentration risks that warrant regulatory and design attention.
• New empirical questions for researchers:
  • How do verified capability signals affect pricing, adoption, and competition between agent creators?
  • What is the elasticity of demand for higher‑guarantee agents versus cheaper, unverifiable options?
  • How does tokenization of agents change incentives for creators, maintenance, and long‑term performance?
  • How will reputation accumulation across tasks translate into measured asset value and returns?
• Policy and regulatory considerations:
  • Persistent linking of creator identities and economic value raises liability, privacy, and compliance questions (consumer protection, KYC/AML where wallets are tied to real creators).
  • Economists and policymakers should examine how this infrastructure interacts with existing legal frameworks for contracts, securities (if agents/tokens qualify), and data protection.
• Practical next steps for empirical work:
  • Prototype deployments and field experiments to measure transaction cost reductions, pricing differentials by verification level, and liquidity in agent asset markets.
  • Laboratory and market simulations to assess strategic behavior under different verification and governance regimes.

Overall, TessIndex outlines an actionable architecture that, if implemented and adopted, could materially alter market structure, contract enforceability, and valuation in the agent economy. For AI economists, this shifts several research fronts from speculative to testable: market design for verified autonomous agents, valuation of persistent agent identities, and regulatory-economics of programmable agent assets.