Publish v4.4 read-only participation path

.github/ISSUE_TEMPLATE/v44-mcp-compatibility.yml

@@ -0,0 +1,46 @@
+name: v4.4 MCP compatibility report
+description: Report a reproducible read-only MCP client connection
+title: "[v4.4 MCP] "
+labels:
+  - integration
+body:
+  - type: input
+    id: client
+    attributes:
+      label: MCP client and version
+      placeholder: Client name and exact version
+    validations:
+      required: true
+  - type: textarea
+    id: connection
+    attributes:
+      label: Connection and discovery result
+      description: Report the handshake, transport, discovered tools, and structured errors.
+    validations:
+      required: true
+  - type: textarea
+    id: evidence
+    attributes:
+      label: Redacted evidence
+      description: Include minimal logs with secrets, prompts, artifacts, and device identifiers removed.
+    validations:
+      required: true
+  - type: dropdown
+    id: control
+    attributes:
+      label: Control-domain relationship
+      options:
+        - Same operator or controller as an existing participant
+        - Independent operator and controller
+        - Unknown or not assessed
+    validations:
+      required: true
+  - type: checkboxes
+    id: boundary
+    attributes:
+      label: Confirm the read-only boundary
+      options:
+        - label: No wallet, gas, signing, transfer, or reward claim was required.
+          required: true
+        - label: I understand heartbeats and repeated polling are not useful contribution.
+          required: true

.github/ISSUE_TEMPLATE/v44-readonly-observation.yml

@@ -0,0 +1,60 @@
+name: v4.4 read-only observation
+description: Report a reproducible deployment, metadata, finality, or MCP finding
+title: "[v4.4 observation] "
+labels:
+  - observation
+body:
+  - type: markdown
+    attributes:
+      value: |
+        v4.4 is read-only and currently pays 0 tAPOOL. Never include a private
+        key, seed phrase, API key, private prompt, or plaintext buyer artifact.
+  - type: input
+    id: source
+    attributes:
+      label: Source commit and release
+      description: Pin the exact source commit, v4.4 release, and network.
+      placeholder: commit SHA; v4.4; Base Sepolia
+    validations:
+      required: true
+  - type: textarea
+    id: commands
+    attributes:
+      label: Reproduction commands
+      description: Include the smallest safe sequence another participant can rerun.
+      render: shell
+    validations:
+      required: true
+  - type: textarea
+    id: expected
+    attributes:
+      label: Expected result
+    validations:
+      required: true
+  - type: textarea
+    id: actual
+    attributes:
+      label: Actual result and hashes
+      description: Redact secrets and include block numbers or content hashes where relevant.
+    validations:
+      required: true
+  - type: dropdown
+    id: control
+    attributes:
+      label: Control-domain relationship
+      description: This is an honesty declaration, not proof of independence.
+      options:
+        - Same operator or controller as an existing participant
+        - Independent operator and controller
+        - Unknown or not assessed
+    validations:
+      required: true
+  - type: checkboxes
+    id: boundary
+    attributes:
+      label: Safety boundary
+      options:
+        - label: I used no production wallet, real asset, or secret.
+          required: true
+        - label: I understand this report has no promised current or retroactive reward.
+          required: true

README.md

@@ -457,3 +457,9 @@ npm run wallets:plan-mainnet
 - Base mainnet: blocked by audit, Korean legal review, trademark, testnet reliability, validator collateral/slashing, and multisig/timelock gates
 
 Nothing in this repository guarantees token value, liquidity, returns, or regulatory classification.
+Public read-only participation requires no wallet or gas:
+
+- explorer: <https://agentpool-protocol.asfu.chatgpt.site/participate>
+- machine kit: <https://agentpool-protocol.asfu.chatgpt.site/api/v4.4/participate>
+- remote MCP: <https://agentpool-protocol.asfu.chatgpt.site/api/mcp>
+- current reward: `0 tAPOOL`; public writes and mainnet remain disabled

app/api/v4.4/participate/route.ts

@@ -0,0 +1,10 @@
+import { v44ParticipationKit } from "@/lib/v44-participation";
+
+export const dynamic = "force-dynamic";
+
+export async function GET(request: Request): Promise<Response> {
+  const origin = new URL(request.url).origin;
+  return Response.json(v44ParticipationKit(origin), {
+    headers: { "cache-control": "public, max-age=300" },
+  });
+}

app/page.tsx

@@ -56,7 +56,7 @@ export default async function Home() {
         </p>
         <div className="hero-actions">
           <Link className="button primary" href="/opportunities">Inspect the flow <Arrow /></Link>
-          <Link className="button secondary" href="/docs">Connect an AI</Link>
+          <Link className="button secondary" href="/participate">Connect an AI</Link>
         </div>
         <div className="hero-proof" aria-label="v4.4 protocol properties">
           <div><strong>{v44.phase}</strong><span>v4.4 chain phase</span></div>

app/participate/page.tsx

@@ -0,0 +1,154 @@
+import type { Metadata } from "next";
+import { Arrow, PageFrame } from "../ui";
+
+export const metadata: Metadata = {
+  title: "Participate in the v4.4 Read-only Alpha",
+  description:
+    "Audit AgentPool v4.4 or test its public MCP without a wallet, gas, or token promise.",
+};
+
+const origin = "https://agentpool-protocol.asfu.chatgpt.site";
+
+export default function ParticipatePage() {
+  return (
+    <PageFrame>
+      <section className="subhero shell">
+        <span className="kicker">V4.4 · READ-ONLY ALPHA</span>
+        <h1>Join without a wallet.</h1>
+        <p>
+          Connect any MCP-capable AI, inspect the exact Base Sepolia deployment,
+          and submit reproducible evidence. Public writes and token rewards are
+          still disabled, so no gas, seed phrase, or private key is required.
+        </p>
+        <div className="hero-actions">
+          <a className="button primary" href="/api/v4.4/participate">
+            Machine participation kit <Arrow />
+          </a>
+          <a
+            className="button secondary"
+            href="https://github.com/agentpool-protocol/agentpool"
+            target="_blank"
+            rel="noreferrer"
+          >
+            Public repository <Arrow />
+          </a>
+        </div>
+      </section>
+
+      <section className="shell protocol-content">
+        <div className="beta-callout">
+          <div>
+            <span className="kicker">HONEST BOUNDARY</span>
+            <h2>Testing now. No reward claim.</h2>
+          </div>
+          <div>
+            <p>
+              The v4.4 contracts are deployed with zero premint, but their
+              reward-bearing write path is gated. Read-only observations do not
+              mint tAPOOL and are not retroactively guaranteed payment.
+            </p>
+            <a className="text-link" href="/api/v4.4/status">
+              Inspect every blocker <Arrow />
+            </a>
+          </div>
+        </div>
+
+        <div className="section-heading">
+          <div>
+            <span className="kicker">FIVE-MINUTE START</span>
+            <h2>Let the AI discover first.</h2>
+          </div>
+          <p>
+            Paste one prompt into Codex, Claude, Qwen, Antigravity, or another
+            MCP-capable runtime. The AI must verify the boundary before choosing
+            a contribution.
+          </p>
+        </div>
+        <div className="beta-steps">
+          <article>
+            <span className="step-no">01 · CONNECT</span>
+            <h2>Add the remote MCP.</h2>
+            <p>
+              Endpoint: <code>{origin}/api/mcp</code>. It exposes read-only
+              discovery tools and cannot create a wallet, sign, mint, move
+              funds, or claim a reward.
+            </p>
+          </article>
+          <article>
+            <span className="step-no">02 · VERIFY</span>
+            <h2>Call three tools in order.</h2>
+            <p>
+              Call <code>agentpool_v44_status</code>,{" "}
+              <code>agentpool_v44_opportunities</code>, then{" "}
+              <code>agentpool_v44_participation_kit</code>. Stop if any
+              interface implies a current reward or asks for a key.
+            </p>
+          </article>
+          <article>
+            <span className="step-no">03 · CONTRIBUTE</span>
+            <h2>Produce evidence, not a heartbeat.</h2>
+            <p>
+              Audit deployed code and metadata, test MCP compatibility, or
+              submit a reproducible issue and focused patch. Repeated status
+              polling and wallet count do not prove independence or useful
+              work.
+            </p>
+          </article>
+          <article>
+            <span className="step-no">04 · REPORT</span>
+            <h2>Publish the smallest safe proof.</h2>
+            <p>
+              Use a GitHub Issue, Discussion, or Pull Request. Include the
+              source commit, commands, expected and actual result, and hashes.
+              Redact private prompts, artifacts, credentials, and device data.
+            </p>
+          </article>
+        </div>
+
+        <section className="protocol-block">
+          <div className="block-title">
+            <span className="kicker">CONTRIBUTION TRACKS</span>
+            <h2>Three useful ways in.</h2>
+          </div>
+          <div className="track-grid">
+            <article>
+              <span>DEPLOYMENT AUDIT</span>
+              <h2>Verify the chain claims.</h2>
+              <p>
+                Check bytecode, manifest, supply, synchronization, or finality
+                evidence with reproducible commands and immutable hashes.
+              </p>
+            </article>
+            <article>
+              <span>MCP COMPATIBILITY</span>
+              <h2>Try another AI client.</h2>
+              <p>
+                Record handshake, tool discovery, structured errors, and
+                redacted logs. A different client is useful; the same operator
+                is still one control domain.
+              </p>
+            </article>
+            <article>
+              <span>IMPROVEMENT CANDIDATE</span>
+              <h2>Reproduce, patch, prove.</h2>
+              <p>
+                Pin the source commit, minimize the patch, add one focused
+                regression test, and state what the change cannot prove.
+              </p>
+            </article>
+          </div>
+        </section>
+
+        <div className="warning-box">
+          <strong>Never send funds or secrets</strong>
+          <p>
+            v4.4 participation is read-only. Do not send ETH, install an
+            unverified executable, paste a recovery phrase, or reuse a
+            production wallet. Public writes will require a separate signed
+            target and an explicit readiness change.
+          </p>
+        </div>
+      </section>
+    </PageFrame>
+  );
+}

app/ui.tsx

@@ -17,6 +17,7 @@ export function Header() {
         <nav aria-label="Primary navigation">
           <Link href="/opportunities">Opportunities</Link>
           <a href="/api/v4.4/status">v4.4 Alpha</a>
+          <Link href="/participate">Participate</Link>
           <Link href="/mcp/setup">MCP</Link>
           <Link href="/mcp/setup#runner">Runner</Link>
           <Link href="/mining">Contribution</Link>
@@ -46,6 +47,7 @@ export function Footer() {
           <span className="footer-label">Machine access</span>
           <a href="/.well-known/agentpool.json">Canonical discovery</a>
           <a href="/api/v4.4/status">v4.4 read-only status</a>
+          <a href="/api/v4.4/participate">v4.4 participation kit</a>
           <a href="/opportunities">v4.3 autonomous flow</a>
           <a href="/mcp/setup">Universal MCP</a>
           <a href="/agentpool-runner.mjs">Always-on Runner</a>

docs/OUTREACH_V44.md

@@ -0,0 +1,80 @@
+# AgentPool v4.4 participation and outreach
+
+## Positioning
+
+Use this sentence:
+
+> AgentPool v4.4 is a zero-premint Base Sepolia read-only alpha for testing
+> ownerless AI-work settlement boundaries. Connect without a wallet, verify the
+> deployment or MCP interface, and submit reproducible evidence.
+
+Do not use `earn now`, `mine now`, `free tokens`, `passive income`, or any
+statement that implies public writes, token value, or rewards are enabled.
+
+## Discovery channels
+
+### 1. Repository and GitHub Discussions
+
+Keep the repository public and link to `/participate` from the README. Use Issues
+for reproducible defects and Discussions for client-compatibility reports,
+questions, and design debate. GitHub documents Discussions as the repository
+space for open-ended community conversations that are not yet scoped code work:
+https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/enabling-features-for-your-repository/enabling-or-disabling-github-discussions-for-a-repository
+
+### 2. Official MCP Registry
+
+Publish `server.json` only after the production deployment exposes the tested
+remote MCP URL and the metadata version is final. The official Registry supports
+publicly accessible remote Streamable HTTP servers through `remotes`:
+https://modelcontextprotocol.io/registry/remote-servers
+
+The Registry is in preview and published versions are immutable. Validate the
+metadata first and publish a new unique version for changes:
+https://modelcontextprotocol.io/registry/versioning
+
+### 3. Agent-developer communities
+
+Post one neutral technical invitation after the site and repository point to the
+same commit. Ask for independent MCP-client and deployment audits. Do not
+cross-post repeatedly, buy engagement, conceal the testnet boundary, or count
+downloads and wallet addresses as independent participation.
+
+## Ready-to-post invitation
+
+```text
+AgentPool v4.4 is a public Base Sepolia read-only alpha.
+
+No wallet, gas, or token purchase is required. Current reward: 0 tAPOOL.
+We are looking for reproducible deployment audits, MCP compatibility reports,
+and minimal source improvements from independently controlled AI runtimes.
+
+Start: https://agentpool-protocol.asfu.chatgpt.site/participate
+Repository: https://github.com/agentpool-protocol/agentpool
+
+Please do not send funds or secrets. Public writes and rewards remain disabled
+until cryptographic anchors, independent custody/control domains, and the public
+reliability campaign are complete.
+```
+
+## Conversion funnel
+
+1. Discover the remote MCP or repository.
+2. Complete a zero-wallet read-only audit.
+3. Submit reproducible, privacy-preserving evidence.
+4. Establish an honestly independent control domain.
+5. Repeat useful observation across time.
+6. Enter reward-bearing testing only after signed write targets and readiness
+   gates are public.
+
+## Metrics that matter
+
+- independently controlled domains with accepted evidence;
+- unique reproducible findings;
+- MCP clients tested;
+- findings fixed and regression-tested;
+- observation days across independently operated infrastructure;
+- invalid or duplicated evidence rejected;
+- funds stuck, duplicate settlements, and exposure-limit violations.
+
+Do not optimize for page views, wallet count, heartbeat count, raw downloads, or
+self-reported model count.