# Kirke Labs

> Kirke Labs is a systems engineering firm building high-reliability software for the next generation of finance and digital ownership. We engineer self-executing, audit-ready systems on Algorand for programmable markets.

## Core
- [Home](https://www.kirkelabs.com/): what Kirke Labs builds, deployments, infrastructure.
- [About](https://www.kirkelabs.com/about): company background, mission, and location.
- [Team](https://www.kirkelabs.com/team): Steve Kirton (Founder & CEO), Soleman El Gelawi (CTO).

## Proof Points
- Top 10 — 2025 Algorand Startup Challenge (Casalotto): https://algorand.co/blog/announcing-the-2025-algorand-startup-challenges-winners
- Casalotto — on-chain prize distribution, top-10 Algorand Startup Challenge 2025: https://www.casalotto.app
- The Marketing Machine — agentic distribution infrastructure, developed in the context of the alliance.xyz accelerator program: https://www.themarketingmachine.app
- agent-readiness-scan — open-source CLI scoring how ready a website is for AI agents across 8 weighted dimensions (discovery, trust, agent-actionability). Intro: https://www.kirkelabs.com/agent-readiness-scan · source: https://github.com/KirkeLabs/agent-readiness-scan
- ai-legibility-scan — open-source CLI scoring how readable a website is to AI agents and LLMs. Intro: https://www.kirkelabs.com/ai-legibility-scan · source: https://github.com/KirkeLabs/ai-legibility-scan
- oaa-agent-kit — open-source (MIT) npm library + CLI for building fundable Algorand agents: a LogicSig mandate that bounds spending, x402 payments, an Open Agent Access passport, and Pera Wallet interop. Intro: https://www.kirkelabs.com/oaa-agent-kit · npm: https://www.npmjs.com/package/@kirkelabs/oaa-agent-kit · source: https://github.com/KirkeLabs/oaa-agent-kit
- walletless-kit — open-source (MIT) npm toolkit for walletless commerce: ephemeral custodial onboarding, receipt-only on-chain proofs, a tamper-evident RFC 6962 Merkle audit trail, segregated money ledgers, and a verifiable, recomputable prize draw with a zero-dependency proof bundle (walletless-proof/v1). The intro page embeds a browser-side verifier that recomputes a sample draw. Intro: https://www.kirkelabs.com/walletless-kit · npm: https://www.npmjs.com/package/@kirkelabs/walletless-kit · source: https://github.com/KirkeLabs/walletless-kit
- Atelier OpenLedger — flagship build: MiCA/PSD2-compliant blockchain-verified charitable procurement; proof on-chain, funds in the charity's own bank. v5.2 whitepaper on request (contact@kirkelabs.com): https://www.atelier-openledger.com
- Agentic Web Readiness proof surface — public dashboard and evidence bundle showing the site's policy discovery, allowed read path, denied training path, compute-heavy review path, TestNet payment posture, receipt example, and scanner-output links: https://www.kirkelabs.com/agentic-web-readiness · bundle: https://www.kirkelabs.com/proof/agentic-web-readiness-evidence.json
- Kirke Labs package integrity matrix — npm/GitHub/provenance snapshot for the public packages and known metadata gaps: https://www.kirkelabs.com/proof/kirke-package-integrity.json · summary: https://www.kirkelabs.com/proof/kirke-package-integrity.md

## Roadmap
- Aurelius Finance — asset-backed finance, awaiting regulatory clarity.

## Optional
- [Tooling](https://www.kirkelabs.com/tooling): Open Agent Access and the agentic-payments stack — x402 settlement on Algorand, walletless onboarding, policy wallets.
- [Ecosystem](https://www.kirkelabs.com/ecosystem): walletless Web3 demo for fractional payouts and atomic refunds.
- [Agent City](https://www.kirkelabs.com/agent-city): interactive simulation of autonomous agent missions.
- [Agent-Readiness Scan](https://www.kirkelabs.com/agent-readiness-scan): intro to the open-source agent-readiness-scan CLI — what it scores, how to run it, and our own 29 → 73 dogfooding story.
- [AI Visibility](https://www.kirkelabs.com/ai-legibility-scan): measure your AI-legibility with the open-source ai-legibility-scan CLI, then fix it with the Prove-It Protocol field guide.
- [Open Agent Access](https://www.kirkelabs.com/open-agent-access): drop-in governance for any MCP server — every tool call is authorised (allow / deny / charge / review), fails closed, and leaves a signed, hash-chained receipt that rolls up into an independently-verifiable evidence bundle.
- [MCP Governance](https://www.kirkelabs.com/mcp-governance): the open-source reference for governing any MCP server with verifiable signed receipts — authorise → fail closed → signed receipt → evidence bundle, in a sub-30-minute quickstart. Source: https://github.com/KirkeLabs/oaa-mcp-governance
- [OAA Agent Kit](https://www.kirkelabs.com/oaa-agent-kit): open-source npm library + CLI for building fundable Algorand agents — LogicSig mandate, x402 payments, OAA passport, Pera Wallet interop.
- [Agentic Web Readiness](https://www.kirkelabs.com/agentic-web-readiness): the proof surface for agents and humans — policy, preflight, allow / deny / charge / review, scanner evidence, and downloadable bundle.
- [Walletless Kit](https://www.kirkelabs.com/walletless-kit): open-source npm toolkit for walletless commerce — ephemeral custodial onboarding, non-PII receipts, a tamper-evident Merkle audit trail, segregated ledgers, and a verifiable, recomputable draw you can re-check in your own browser.

## Writing
- [The Prove-It Protocol](https://www.stevekirton.com): a 10-step field guide by Steve Kirton (founder, Kirke Labs) on being found by AI — engineering visibility, trust, structured data, and entity SEO in the interpretation economy. Full guide on stevekirton.com; summarised on the AI Visibility page at https://www.kirkelabs.com/ai-legibility-scan.

## Agents (Open Agent Access)
What this demonstrates: consequential agents must carry papers. Agent City (https://www.kirkelabs.com/agent-city) is the visible simulation; Open Agent Access is the control layer underneath — identity, mandate, purpose, policy decision, budget, payment terms, receipt, and provenance for every agent action. allow / deny / charge decisions and the hash-chained receipts are produced by Kirke Labs' live policy engine, with payment terms settled via x402 on Algorand. This is the infrastructure Kirke Labs builds: governance and payments for autonomous commerce.
This site publishes an Open Agent Access policy. Agents: read it, send a passport (AA-Agent-ID, AA-Purpose, AA-Use), respect the decision, and attribute by source URL. Public reads are allowed; AI-training use is denied; paid actions return an x402 charge (402).
- [Agent Access policy](https://www.kirkelabs.com/.well-known/agent-access.json): identity, rules, rate limits, contacts.
- [Agent City scenarios](https://www.kirkelabs.com/agent-city-scenarios.json): machine-readable scenario registry.
- [Agentic Web Readiness evidence bundle](https://www.kirkelabs.com/proof/agentic-web-readiness-evidence.json): governed proof manifest with exact simulated/TestNet/production labels.
- [Receipt example](https://www.kirkelabs.com/proof/agentic-web-readiness-receipt.json): unsigned example receipt for the public proof route.
- API: POST /api/agent-city/preflight (evaluate), POST /api/agent-city/step (execute, passport required), GET /api/agent-city/policy, GET /api/agent-city/scenarios.

Escalation contact for denied or reviewed agent requests: contact@kirkelabs.com. Security contact: security@kirkelabs.com.

## Discovery (well-known)
- /.well-known/agent-access.json — Open Agent Access policy (identity, rules, payment terms).
- /.well-known/ucp — agentic-commerce manifest (Universal Commerce Protocol), settled via x402 on Algorand.
- /.well-known/security.txt — security contact (RFC 9116).
- /.well-known/http-message-signatures-directory — Web Bot Auth public key (Ed25519; forward-looking).
- /agent-city-scenarios.json — machine-readable scenario registry.
- robots.txt carries Cloudflare Content Signals: search=yes, ai-input=yes, ai-train=no.

## Legal
- [Privacy Policy](https://www.kirkelabs.com/privacy): GDPR contact and rights; no accounts or ad trackers.
- [Terms of Service](https://www.kirkelabs.com/terms): informational/demo use; Agent City runs on Algorand TestNet.