LeoYeAI/agent-master-skills / skills/agent-comm-hub
agent-comm-hub
Flagged多智能体协同通信基础设施——基于 MCP+SSE 的实时消息、任务调度、记忆共享与进化引擎。支持 WorkBuddy、Hermes、QClaw 及任意 MCP 兼容 Agent 接入。53 个 MCP 工具、4 级权限、零外部依赖 Python SDK。触发词:agent通信、智能体通信、hub通信、多智能体、跨agent通信、任务调度、assign_task、send_message、hermes通信、workbuddy通信、agent hub、通信hub、mcp通信、记忆共享、进化引擎、策略共享、经验分享、共享记忆,共同进化
src/tools/identity.ts
View on GitHub/**
* identity.ts — MCP 工具:Agent Identity 模块
* 包含:register_agent, heartbeat, query_agents, get_online_agents, set_trust_score, revoke_token
* 来源:tools.ts 第 125-319 行 + 第 854-873 行
*/
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { z } from "zod";
import {
checkPermission,
getRequiredPermission,
auditLog,
revokeToken as revokeTokenFromSecurity,
recalculateTrustScore,
type AuthContext,
} from "../security.js";
import {
registerAgent as registerAgentFromIdentity,
heartbeat as heartbeatFromIdentity,
queryAgents as queryAgentsFromIdentity,
clearOfflineNotification,
updateAgentTrustScore,
resolveAgentId,
} from "../identity.js";
import { onlineAgents } from "../sse.js";
import { db } from "../db.js";
import { withRetry, requireAuth, mcpFail } from "../utils.js";
export function registerIdentityTools(server: McpServer, authContext?: AuthContext): void {
// ────────────────────────────────────────────────────
// NEW Tool 1: register_agent (Phase 1)
// 注册新 Agent — public 工具,无需认证
// ────────────────────────────────────────────────────
server.tool(
"register_agent",
"注册新 Agent 到 Hub。需要有效的邀请码。注册成功返回 agent_id 和 api_token(仅显示一次)。",
{
invite_code: z.string().describe("邀请码(通过 /admin/invite/generate 获取)"),
name: z.string().describe("Agent 名称"),
capabilities: z.array(z.string()).optional()
.describe("Agent 能力列表,如 ['mcp', 'sse', 'memory']"),
},
async ({ invite_code, name, capabilities }) => {
// public 工具,不需要权限检查
const result = registerAgentFromIdentity(invite_code, name, capabilities || []);
if (!result.success) {
return mcpFail(result.error ?? "Registration failed", "register_agent");
}
auditLog("tool_register_agent", result.agentId ?? null, name, `role=${result.role ?? 'unknown'}`);
return {
content: [{
type: "text",
text: JSON.stringify({
success: true,
agent_id: result.agentId,
api_token: result.apiToken,
role: result.role ?? "member",
warning: "⚠️ api_token 仅显示一次,请妥善保存!",
next_step: "使用此 Token 调用 heartbeat 工具上线,Token 通过 Authorization: Bearer <token> 传递",
}, null, 2),
}],
};
}
);
// ────────────────────────────────────────────────────
// NEW Tool 2: heartbeat (Phase 1)
// 上报心跳 — member 及以上
// ────────────────────────────────────────────────────
server.tool(
"heartbeat",
"上报 Agent 心跳,维持在线状态并累积信任分。Agent 上线后应每 30 秒调用一次。超过 90 秒无心跳将自动标记为离线。连续在线心跳每 3 次自动增加 1 点 trust_score(上限 100)。",
{
agent_id: z.string().describe("Agent ID(注册时返回的 agent_id)"),
},
async ({ agent_id }) => {
const ctx = requireAuth(authContext, "heartbeat");
// 验证调用者是 Agent 本人
if (ctx.agentId !== agent_id && ctx.role !== "admin") {
return {
content: [{
type: "text",
text: JSON.stringify({
success: false,
error: "Cannot send heartbeat for another agent (admin only)",
}),
}],
};
}
const result = heartbeatFromIdentity(agent_id);
if (result.success) {
// 清除离线通知标记
clearOfflineNotification(agent_id);
}
return {
content: [{
type: "text",
text: JSON.stringify(result, null, 2),
}],
};
}
);
// ────────────────────────────────────────────────────
// NEW Tool 3: query_agents (Phase 1)
// 查询已注册 Agent — member 及以上
// ────────────────────────────────────────────────────
server.tool(
"query_agents",
"查询已注册的 Agent 列表。支持按状态、角色筛选。",
{
status: z.enum(["online", "offline", "all"]).optional()
.default("all").describe("Agent 状态筛选"),
role: z.enum(["admin", "member", "group_admin"]).optional()
.describe("角色筛选"),
capability: z.string().optional()
.describe("能力筛选"),
},
async ({ status, role, capability }) => {
const ctx = requireAuth(authContext, "query_agents");
const agents = queryAgentsFromIdentity({ status, role, capability });
return {
content: [{
type: "text",
text: JSON.stringify({
agents,
count: agents.length,
queried_by: ctx.agentId,
}, null, 2),
}],
};
}
);
// ────────────────────────────────────────────────────
// NEW Tool 4: revoke_token (Phase 1)
// 吊销 Token — admin only
// ────────────────────────────────────────────────────
server.tool(
"revoke_token",
"吊销 API Token,使其立即失效。仅 admin 可调用。",
{
token_id: z.string().describe("要吊销的 Token ID"),
},
async ({ token_id }) => {
const ctx = requireAuth(authContext, "revoke_token");
const success = revokeTokenFromSecurity(token_id);
if (success) {
auditLog("tool_revoke_token", ctx.agentId, token_id);
// Phase 5a Day 2: token 吊销影响信任评分
try {
const tokenRow = db.prepare(`SELECT agent_id FROM auth_tokens WHERE token_id=?`).get(token_id) as any;
if (tokenRow?.agent_id) {
recalculateTrustScore(tokenRow.agent_id);
}
} catch {}
}
return {
content: [{
type: "text",
text: JSON.stringify({
success,
token_id,
note: success ? "Token 已吊销,使用该 Token 的 Agent 将无法访问" : "Token not found or already revoked",
}, null, 2),
}],
};
}
);
// ────────────────────────────────────────────────────
// NEW Tool 4b: set_trust_score (Phase 2 Day 4)
// 设置信任分 — admin only
// ────────────────────────────────────────────────────
server.tool(
"set_trust_score",
"调整 Agent 信任分(-100 到 +100 的增量)。信任分影响 collective 记忆搜索排序,高信任 Agent 的记忆排名靠前。仅 admin 可调用。",
{
agent_id: z.string().describe("目标 Agent ID"),
delta: z.number().min(-100).max(100).describe("信任分增量(正数加分,负数扣分)"),
},
async ({ agent_id, delta }) => {
const ctx = requireAuth(authContext, "set_trust_score");
const result = updateAgentTrustScore(agent_id, delta, ctx.agentId);
if (!result.ok) {
return mcpFail(result.error, "set_trust_score");
}
return {
content: [{
type: "text",
text: JSON.stringify({
success: true,
agent_id,
new_score: result.new_score,
delta,
note: result.new_score >= 80
? "🟢 高信任 Agent,记忆搜索排名优先"
: result.new_score >= 30
? "🟡 正常信任分"
: "🔴 低信任 Agent,记忆搜索排名靠后",
}, null, 2),
}],
};
}
);
// ────────────────────────────────────────────────────
// Tool 10: get_online_agents (原有,添加权限检查)
// ────────────────────────────────────────────────────
server.tool(
"get_online_agents",
"查询当前通过 SSE 在线连接的 Agent 列表,分配任务前可先确认对方在线。",
{},
async () => {
requireAuth(authContext, "get_online_agents");
const online = onlineAgents();
return {
content: [{
type: "text",
text: JSON.stringify({
online_agents: online,
count: online.length,
timestamp: Date.now(),
}, null, 2),
}],
};
}
);
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |