@clawcrony/claw-crony 1.1.0 → 1.2.3

package/README.md

@@ -1,87 +1,95 @@
-# Claw Crony
-
-OpenClaw A2A v0.3.0 Gateway — Auto-discovery and secure communication between OpenClaw Agents on different servers.
-
-[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
-[![A2A v0.3.0](https://img.shields.io/badge/A2A-v0.3.0-green.svg)](https://github.com/google/A2A)
-
-## Key Features
-
-- **A2A Protocol v0.3.0** — JSON-RPC / REST / gRPC with automatic fallback
-- **Hub Matchmaking** — Auto-match peer Agents by skills with token exchange
-- **Smart Routing** — Auto-select targets by message patterns, tags, or peer skills
-- **Secure Auth** — Bearer Token + zero-downtime multi-token rotation
-- **Resilience** — Health checks + exponential backoff + circuit breaker
-- **File Transfer** — URI / base64 / MIME whitelist + SSRF protection
-- **Observability** — JSONL audit logs + Telemetry metrics endpoint
-
-## Hub Server
-
-Default Hub: `https://www.factormining.cn`
-
-After installation, the plugin auto-registers with the Hub (requires `registrationEnabled: true`). Once registered, use the `a2a_match_request` tool to发起匹配请求, and the Hub will return available peer Agent addresses and auth tokens based on skills.
-
-A2A service port: **18800** (default)
-
-## Installation
-
-### Via npm (Recommended)
-
-```bash
-npm install @clawcrony/claw-crony
-```
-
-### Via Git Clone
-
-```bash
-git clone https://github.com/ccccl8/claw-crony.git
-cd claw-crony
-npm install
-openclaw plugins install .
-openclaw gateway restart
-
-# Verify
-curl -s http://localhost:18800/.well-known/agent-card.json
-```
-
-## Adding a Peer
-
-```bash
-openclaw config set plugins.entries.claw-crony.config.peers '[{
-  "name": "Peer Name",
-  "agentCardUrl": "http://<peerIP>:18800/.well-known/agent-card.json",
-  "auth": { "type": "bearer", "token": "<peerToken>" }
-}]'
-openclaw gateway restart
-```
-
-## Hub Matchmaking (a2a_match_request)
-
-发起匹配请求 to the Hub, which automatically finds registered Agents with the required skills:
-
-```bash
-# Agent calls a2a_match_request tool with params:
-# { skills: ["chat"], description?: "optional description" }
-#
-# Returns: provider address + yourToken + peerToken
-# Both sides configure each other as peers using the returned tokens to communicate
-```
-
-For detailed configuration steps, see [CONFIG.md](CONFIG.md).
-
-## Endpoints
-
-| Endpoint | Method | Description |
-|----------|--------|-------------|
-| `/.well-known/agent-card.json` | GET | Agent Card (discovery) |
-| `/a2a/jsonrpc` | POST | A2A JSON-RPC |
-| `/a2a/rest` | POST | A2A REST transport |
-| `/a2a/metrics` | GET | Telemetry snapshot (when enabled) |
-
-## License
-
-MIT License
-
-## Acknowledgments
-
-This project is based on [win4r/openclaw-a2a-gateway](https://github.com/win4r/openclaw-a2a-gateway), MIT License.
+# Claw Crony
+
+OpenClaw A2A v0.3.0 Gateway - Auto-discovery and secure communication between OpenClaw Agents on different servers.
+
+[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
+[![A2A v0.3.0](https://img.shields.io/badge/A2A-v0.3.0-green.svg)](https://github.com/google/A2A)
+
+## Key Features
+
+- **A2A Protocol v0.3.0** - JSON-RPC / REST / gRPC with automatic fallback
+- **Hub Matchmaking** - Auto-match peer Agents by skills with token exchange
+- **Smart Routing** - Auto-select targets by message patterns, tags, or peer skills
+- **Secure Auth** - Bearer Token + zero-downtime multi-token rotation
+- **Resilience** - Health checks + exponential backoff + circuit breaker
+- **File Transfer** - URI / base64 / MIME whitelist + SSRF protection
+- **Observability** - JSONL audit logs + Telemetry metrics endpoint
+
+## Hub Server
+
+Default Hub: `https://www.clawcrony.com`
+
+After installation, the plugin auto-registers with the Hub (requires `registrationEnabled: true`). Once registered, use the `a2a_match_request` tool to send a matchmaking request, and the Hub will return a matched peer Agent address together with the tokens needed for the current session.
+
+After the user signs in to the Hub web dashboard, they can currently see:
+
+- Their own Agent profile, address, and normalized skill tags
+- A match timeline for requests created by this Agent
+- Per-request request summary, required skills, and current status
+- Matched result details including provider name, provider address, and update time
+- Whether requester/provider tokens have already been submitted for the match
+
+A2A service port: **18800** (default)
+
+## Installation
+
+### Via npm (Recommended)
+
+```bash
+npm install @clawcrony/claw-crony
+```
+
+### Via Git Clone
+
+```bash
+git clone https://github.com/ccccl8/claw-crony.git
+cd claw-crony
+npm install
+openclaw plugins install .
+openclaw gateway restart
+
+# Verify
+curl -s http://localhost:18800/.well-known/agent-card.json
+```
+
+## Adding a Peer
+
+```bash
+openclaw config set plugins.entries.claw-crony.config.peers '[{
+  "name": "Peer Name",
+  "agentCardUrl": "http://<peerIP>:18800/.well-known/agent-card.json",
+  "auth": { "type": "bearer", "token": "<peerToken>" }
+}]'
+openclaw gateway restart
+```
+
+## Hub Matchmaking (a2a_match_request)
+
+Send a matchmaking request to the Hub, which automatically finds registered Agents with the required skills:
+
+```bash
+# Agent calls a2a_match_request tool with params:
+# { skills: ["chat"], description?: "optional description" }
+#
+# Returns: provider address + yourToken + peerToken
+# Both sides configure each other as peers using the returned tokens to communicate
+```
+
+For detailed configuration steps, see [CONFIG.md](CONFIG.md).
+
+## Endpoints
+
+| Endpoint | Method | Description |
+|----------|--------|-------------|
+| `/.well-known/agent-card.json` | GET | Agent Card (discovery) |
+| `/a2a/jsonrpc` | POST | A2A JSON-RPC |
+| `/a2a/rest` | POST | A2A REST transport |
+| `/a2a/metrics` | GET | Telemetry snapshot (when enabled) |
+
+## License
+
+MIT License
+
+## Acknowledgments
+
+This project is based on [win4r/openclaw-a2a-gateway](https://github.com/win4r/openclaw-a2a-gateway), MIT License.

package/dist/index.js

@@ -24,6 +24,7 @@ import { AuditLogger } from "./src/audit.js";
 import { PeerHealthManager } from "./src/peer-health.js";
 import { runHubRegistration } from "./src/hub-registration.js";
 import { HubMatchClient } from "./src/hub-match.js";
+import { normalizeAgentCardSkills } from "./src/skill-catalog.js";
 import { parseRoutingRules, matchRule } from "./src/routing-rules.js";
 import { validateUri, validateMimeType, } from "./src/file-security.js";
 /** Build a JSON-RPC error response. */
@@ -71,7 +72,7 @@ function extractSkillsFromAgentCard(card) {
     return skills.map((s) => (typeof s === "string" ? s : asObject(s).name ?? "")).filter(Boolean);
 }
 function parseAgentCard(raw) {
-    const skills = Array.isArray(raw.skills) ? raw.skills : [];
+    const skills = normalizeAgentCardSkills(Array.isArray(raw.skills) ? raw.skills : []);
     return {
         name: asString(raw.name, "OpenClaw A2A Gateway"),
         description: asString(raw.description, "A2A bridge for OpenClaw agents"),
@@ -203,7 +204,7 @@ export function parseConfig(raw, resolvePath) {
             },
         },
         hub: {
-            url: asString(hub.url, "https://www.factormining.cn"),
+            url: asString(hub.url, "https://www.clawcrony.com"),
             enabled: asBoolean(hub.enabled, true),
             registrationEnabled: asBoolean(hub.registrationEnabled, true),
         },
@@ -220,6 +221,60 @@ function normalizeCardPath() {
     }
     return `/${AGENT_CARD_PATH}`;
 }
+function getAdvertisedInboundToken(config, hubClient) {
+    if (config.security.tokens && config.security.tokens.length > 0) {
+        return config.security.tokens[0] ?? null;
+    }
+    if (config.security.token) {
+        return config.security.token;
+    }
+    return hubClient?.registrationToken ?? null;
+}
+async function processPendingHubMatches(api, config, processedMatches) {
+    let hubClient;
+    try {
+        hubClient = await HubMatchClient.create();
+    }
+    catch (err) {
+        api.logger.warn(`claw-crony: pending match polling skipped - ${err instanceof Error ? err.message : String(err)}`);
+        return;
+    }
+    const inboundToken = getAdvertisedInboundToken(config, hubClient);
+    if (!inboundToken) {
+        api.logger.warn("claw-crony: pending match polling skipped - no inbound token available");
+        return;
+    }
+    let matches;
+    try {
+        matches = await hubClient.getPendingMatches();
+    }
+    catch (err) {
+        api.logger.warn(`claw-crony: failed to fetch pending matches - ${err instanceof Error ? err.message : String(err)}`);
+        return;
+    }
+    for (const match of matches) {
+        if (match.callerRole !== "provider") {
+            continue;
+        }
+        try {
+            const alreadySubmitted = match.providerTokenSubmitted === true || processedMatches.has(match.id);
+            let currentMatch = match;
+            if (!alreadySubmitted) {
+                currentMatch = await hubClient.submitToken(match.id, inboundToken);
+                processedMatches.add(match.id);
+                api.logger.info(`claw-crony: submitted provider token for hub match ${match.id}`);
+            }
+            if (currentMatch.readyForComplete === true && currentMatch.status === "token_exchange") {
+                await hubClient.completeMatch(match.id, inboundToken);
+                processedMatches.delete(match.id);
+                api.logger.info(`claw-crony: completed hub match ${match.id}`);
+            }
+        }
+        catch (err) {
+            api.logger.warn(`claw-crony: failed to process hub match ${match.id} - ${err instanceof Error ? err.message : String(err)}`);
+        }
+    }
+}
 const plugin = {
     id: "claw-crony",
     name: "Claw Crony",
@@ -340,7 +395,9 @@ const plugin = {
         let server = null;
         let grpcServer = null;
         let cleanupTimer = null;
+        let hubMatchPollingTimer = null;
         const grpcPort = config.server.port + 1;
+        const processedHubMatches = new Set();
         api.registerGatewayMethod("a2a.metrics", ({ respond }) => {
             respond(true, {
                 metrics: telemetry.snapshot(),
@@ -548,7 +605,7 @@ const plugin = {
                         match = await client.createMatch({
                             skills: params.skills,
                             description: params.description,
-                            token: client["registration"].token,
+                            token: getAdvertisedInboundToken(config, client) ?? client.registrationToken,
                         });
                     }
                     catch (err) {
@@ -561,7 +618,7 @@ const plugin = {
                     // Submit our token
                     let updatedMatch;
                     try {
-                        updatedMatch = await client.submitToken(match.id, client["registration"].token);
+                        updatedMatch = await client.submitToken(match.id, getAdvertisedInboundToken(config, client) ?? client.registrationToken);
                     }
                     catch (err) {
                         const msg = err instanceof Error ? err.message : String(err);
@@ -572,24 +629,24 @@ const plugin = {
                     }
                     const provider = updatedMatch.provider;
                     const providerAddress = provider?.address ?? "(unknown)";
-                    const yourToken = updatedMatch.yourToken ?? "(none)";
-                    const peerToken = updatedMatch.peerToken ?? "(none)";
+                    const providerAccessToken = updatedMatch.yourToken ?? "(none)";
+                    const requesterInboundToken = updatedMatch.peerToken ?? "(none)";
                     const status = updatedMatch.status;
                     return {
                         content: [{
                                 type: "text",
                                 text: `Match ${status}: id=${updatedMatch.id}\n` +
                                     `Provider: ${provider?.name ?? "(unknown)"} at ${providerAddress}\n` +
-                                    `Your token (use to contact provider): ${yourToken}\n` +
-                                    `Peer token (provider's token to contact you): ${peerToken}`,
+                                    `Provider access token (use to contact provider): ${providerAccessToken}\n` +
+                                    `Requester inbound token (provider uses this to contact you): ${requesterInboundToken}`,
                             }],
                         details: {
                             ok: true,
                             matchId: updatedMatch.id,
                             status: updatedMatch.status,
                             providerAddress,
-                            yourToken,
-                            peerToken,
+                            yourToken: providerAccessToken,
+                            peerToken: requesterInboundToken,
                         },
                     };
                 },
@@ -682,6 +739,15 @@ const plugin = {
                 // Run once at startup to clear any backlog
                 doCleanup();
                 cleanupTimer = setInterval(doCleanup, intervalMs);
+                if (config.hub?.enabled !== false) {
+                    const pollingIntervalMs = Math.max(5_000, config.resilience.healthCheck.intervalMs);
+                    const pollHubMatches = () => {
+                        void processPendingHubMatches(api, config, processedHubMatches);
+                    };
+                    pollHubMatches();
+                    hubMatchPollingTimer = setInterval(pollHubMatches, pollingIntervalMs);
+                    api.logger.info(`claw-crony: hub match polling enabled interval=${pollingIntervalMs}ms`);
+                }
                 api.logger.info(`claw-crony: task cleanup enabled — ttl=${config.storage.taskTtlHours}h interval=${config.storage.cleanupIntervalMinutes}min`);
             },
             async stop(_ctx) {
@@ -693,6 +759,10 @@ const plugin = {
                     clearInterval(cleanupTimer);
                     cleanupTimer = null;
                 }
+                if (hubMatchPollingTimer) {
+                    clearInterval(hubMatchPollingTimer);
+                    hubMatchPollingTimer = null;
+                }
                 // Stop gRPC server
                 if (grpcServer) {
                     grpcServer.forceShutdown();

package/dist/src/hub-match.d.ts

@@ -18,16 +18,23 @@ export interface HubAgentDto {
 }
 export interface HubMatchResult {
     id: number;
+    requestId?: number | null;
     status: string;
     requester: HubAgentDto | null;
     provider: HubAgentDto | null;
     yourToken: string | null;
     peerToken: string | null;
+    callerRole?: "requester" | "provider" | "observer" | null;
+    requesterTokenSubmitted?: boolean;
+    providerTokenSubmitted?: boolean;
+    readyForComplete?: boolean;
 }
 export declare class HubMatchClient {
     private readonly hubUrl;
     private readonly registration;
     constructor(hubUrl: string, registration: HubRegistrationData);
+    get agentId(): number;
+    get registrationToken(): string;
     static create(): Promise<HubMatchClient>;
     private request;
     /**

package/dist/src/hub-match.js

@@ -20,6 +20,12 @@ export class HubMatchClient {
         this.hubUrl = hubUrl.replace(/\/$/, "");
         this.registration = registration;
     }
+    get agentId() {
+        return this.registration.agentId;
+    }
+    get registrationToken() {
+        return this.registration.token;
+    }
     static async create() {
         const registration = loadRegistration();
         if (!registration) {

package/dist/src/skill-catalog.d.ts

@@ -0,0 +1,4 @@
+import type { AgentSkillConfig } from "./types.js";
+export declare const PRESET_AGENT_SKILLS: readonly ["chat", "search", "reasoning", "tool_use", "file_transfer", "image_understanding", "audio_understanding", "translation", "summarization", "ocr", "code_generation", "code_review", "data_analysis"];
+export declare function normalizeConfiguredSkillName(value: string): string;
+export declare function normalizeAgentCardSkills(rawSkills: Array<AgentSkillConfig | string> | undefined): Array<AgentSkillConfig | string>;

package/dist/src/skill-catalog.js

@@ -0,0 +1,53 @@
+export const PRESET_AGENT_SKILLS = [
+    "chat",
+    "search",
+    "reasoning",
+    "tool_use",
+    "file_transfer",
+    "image_understanding",
+    "audio_understanding",
+    "translation",
+    "summarization",
+    "ocr",
+    "code_generation",
+    "code_review",
+    "data_analysis",
+];
+export function normalizeConfiguredSkillName(value) {
+    return value
+        .trim()
+        .toLowerCase()
+        .replace(/[\s-]+/g, "_")
+        .replace(/_+/g, "_");
+}
+export function normalizeAgentCardSkills(rawSkills) {
+    if (!rawSkills || rawSkills.length === 0) {
+        return [{ id: "chat", name: "chat", description: "Chat bridge" }];
+    }
+    const seen = new Set();
+    const normalized = [];
+    for (const entry of rawSkills) {
+        if (typeof entry === "string") {
+            const normalizedName = normalizeConfiguredSkillName(entry);
+            if (!normalizedName || seen.has(normalizedName)) {
+                continue;
+            }
+            seen.add(normalizedName);
+            normalized.push(normalizedName);
+            continue;
+        }
+        const normalizedName = normalizeConfiguredSkillName(entry.name);
+        if (!normalizedName || seen.has(normalizedName)) {
+            continue;
+        }
+        seen.add(normalizedName);
+        normalized.push({
+            id: entry.id ? normalizeConfiguredSkillName(entry.id) : normalizedName,
+            name: normalizedName,
+            description: entry.description?.trim() || normalizedName,
+        });
+    }
+    return normalized.length > 0
+        ? normalized
+        : [{ id: "chat", name: "chat", description: "Chat bridge" }];
+}

package/openclaw.plugin.json

@@ -2,18 +2,18 @@
   "id": "claw-crony",
   "name": "Claw Crony",
   "description": "OpenClaw A2A v0.3.0 gateway with Agent Card, JSON-RPC, REST, routing rules, transport fallback, and Hub matchmaking",
-  "version": "1.0.1",
+  "version": "1.2.3",
   "defaultConfig": {
     "agentCard": {
       "name": "OpenClaw A2A Gateway",
       "description": "A2A bridge for OpenClaw agents",
       "skills": [{ "id": "chat", "name": "chat", "description": "Chat bridge" }]
-    },
-    "hub": {
-      "url": "https://www.factormining.cn",
-      "enabled": true,
-      "registrationEnabled": true
-    }
+    },
+    "hub": {
+      "url": "https://www.clawcrony.com",
+      "enabled": true,
+      "registrationEnabled": true
+    }
   },
   "configSchema": {
     "type": "object",
@@ -26,13 +26,14 @@
           "name": { "type": "string" },
           "description": { "type": "string" },
           "url": { "type": "string" },
-          "skills": {
-            "type": "array",
-            "items": {
-              "oneOf": [
-                { "type": "string" },
-                {
-                  "type": "object",
+          "skills": {
+            "type": "array",
+            "description": "Agent skills sent to the Hub. Prefer preset names such as chat, search, reasoning, tool_use, file_transfer, image_understanding, audio_understanding, translation, summarization, ocr, code_generation, code_review, data_analysis. Custom skills are also allowed.",
+            "items": {
+              "oneOf": [
+                { "type": "string" },
+                {
+                  "type": "object",
                   "properties": {
                     "id": { "type": "string" },
                     "name": { "type": "string" },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@clawcrony/claw-crony",
-  "version": "1.1.0",
+  "version": "1.2.3",
   "type": "module",
   "description": "OpenClaw A2A gateway plugin implementing the A2A v0.3.0 protocol surface",
   "main": "dist/index.js",