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/repo/sqlite-impl.ts
View on GitHub/**
* repo/sqlite-impl.ts — IMessageRepo / ITaskRepo / IConsumedLogRepo 的 SQLite 实现
* Phase 2 Day 2 基础 + Phase 4b 依赖链 + 质量门扩展
*/
import {
db,
msgStmt,
taskStmt,
consumedStmt,
type Message,
type Task,
type ConsumedEntry,
} from "../db.js";
import type { IMessageRepo, ITaskRepo, IConsumedLogRepo } from "./interfaces.js";
import type { TaskDependency, QualityGate, DepType, DepStatus, GateStatus } from "./types.js";
// ─── MessageRepo ──────────────────────────────────────────────
class SqliteMessageRepo implements IMessageRepo {
insert(msg: Message): void {
msgStmt.insert.run(msg);
}
markDelivered(id: string): void {
msgStmt.markDelivered.run(id);
}
markRead(id: string): void {
msgStmt.markRead.run(id);
}
markAcknowledged(id: string): void {
msgStmt.markAcknowledged.run(id);
}
markAllDelivered(toAgent: string): void {
msgStmt.markAllDelivered.run(toAgent);
}
pendingFor(toAgent: string): Message[] {
return msgStmt.pendingFor.all(toAgent) as Message[];
}
getById(id: string): Message | undefined {
return msgStmt.getById.get(id) as Message | undefined;
}
listByStatus(toAgent: string, status: string): Message[] {
const stmt = db.prepare(
`SELECT * FROM messages WHERE to_agent=? AND status=? ORDER BY created_at ASC`
);
return stmt.all(toAgent, status) as Message[];
}
updateStatus(id: string, status: string): void {
const stmt = db.prepare(`UPDATE messages SET status=? WHERE id=?`);
stmt.run(status, id);
}
listSince(toAgent: string, since: number): Message[] {
const stmt = db.prepare(
`SELECT * FROM messages WHERE to_agent=? AND created_at > ? ORDER BY created_at ASC`
);
return stmt.all(toAgent, since) as Message[];
}
}
// ─── TaskRepo ─────────────────────────────────────────────────
class SqliteTaskRepo implements ITaskRepo {
insert(task: Task): void {
taskStmt.insert.run(task);
}
getById(id: string): Task | undefined {
return taskStmt.getById.get(id) as Task | undefined;
}
update(id: string, status: string, result: string | null, progress: number): void {
taskStmt.update.run(status, result, progress, Date.now(), id);
}
assignTo(id: string, assignedTo: string): void {
taskStmt.updateAssignee.run(assignedTo, Date.now(), Date.now(), id);
}
listFor(assignedTo: string, status: string): Task[] {
return taskStmt.listFor.all(assignedTo, status) as Task[];
}
listByPipeline(pipelineId: string): Task[] {
return taskStmt.listByPipeline.all(pipelineId) as Task[];
}
// ─── Phase 4b: 依赖链实现 ────────────────────────────────
addDependency(upstreamId: string, downstreamId: string, depType: DepType = "finish_to_start"): TaskDependency {
if (upstreamId === downstreamId) {
throw new Error("Cannot create self-dependency");
}
// 检查是否已存在
const existing = db.prepare(
`SELECT * FROM task_dependencies WHERE upstream_id=? AND downstream_id=?`
).get(upstreamId, downstreamId);
if (existing) {
throw new Error(`Dependency already exists: ${upstreamId} → ${downstreamId}`);
}
const id = `dep_${Date.now()}_${Math.random().toString(36).slice(2, 8)}`;
const now = Date.now();
const dep: TaskDependency = {
id, upstream_id: upstreamId, downstream_id: downstreamId,
dep_type: depType, status: "pending", created_at: now,
};
db.prepare(
`INSERT INTO task_dependencies (id, upstream_id, downstream_id, dep_type, status, created_at)
VALUES (@id, @upstream_id, @downstream_id, @dep_type, @status, @created_at)`
).run(dep);
return dep;
}
removeDependency(upstreamId: string, downstreamId: string): void {
const result = db.prepare(
`DELETE FROM task_dependencies WHERE upstream_id=? AND downstream_id=?`
).run(upstreamId, downstreamId);
if (result.changes === 0) {
throw new Error(`Dependency not found: ${upstreamId} → ${downstreamId}`);
}
}
getDependencies(taskId: string): { upstreams: TaskDependency[]; downstreams: TaskDependency[] } {
const upstreams = db.prepare(
`SELECT * FROM task_dependencies WHERE downstream_id=? ORDER BY created_at ASC`
).all(taskId) as TaskDependency[];
const downstreams = db.prepare(
`SELECT * FROM task_dependencies WHERE upstream_id=? ORDER BY created_at ASC`
).all(taskId) as TaskDependency[];
return { upstreams, downstreams };
}
checkDependenciesSatisfied(taskId: string): boolean {
// 查找所有 pending 状态的上游依赖
const pending = db.prepare(
`SELECT COUNT(*) as cnt FROM task_dependencies
WHERE downstream_id=? AND status='pending'`
).get(taskId) as { cnt: number };
return pending.cnt === 0;
}
setTaskWaiting(taskId: string): void {
db.prepare(
`UPDATE tasks SET status='waiting', updated_at=? WHERE id=?`
).run(Date.now(), taskId);
}
setTaskReady(taskId: string): void {
const task = db.prepare(`SELECT * FROM tasks WHERE id=?`).get(taskId) as Task | undefined;
if (!task) throw new Error(`Task not found: ${taskId}`);
if (task.status !== "waiting") {
throw new Error(`Task is not in waiting state: ${task.status}`);
}
db.prepare(
`UPDATE tasks SET status='assigned', updated_at=? WHERE id=?`
).run(Date.now(), taskId);
}
wouldCreateCycle(upstreamId: string, downstreamId: string): boolean {
// DFS: 从 downstreamId 出发,沿现有依赖的 downstream 方向搜索,
// 看是否能到达 upstreamId。
// 如果能到达,说明添加 upstreamId→downstreamId 后会形成环。
const visited = new Set<string>();
const stack = [downstreamId];
while (stack.length > 0) {
const current = stack.pop()!;
if (current === upstreamId) return true;
if (visited.has(current)) continue;
visited.add(current);
const deps = db.prepare(
`SELECT downstream_id FROM task_dependencies WHERE upstream_id=?`
).all(current) as Array<{ downstream_id: string }>;
for (const d of deps) {
stack.push(d.downstream_id);
}
}
return false;
}
satisfyDownstream(upstreamId: string): number {
const result = db.prepare(
`UPDATE task_dependencies SET status='satisfied'
WHERE upstream_id=? AND status='pending' AND dep_type='finish_to_start'`
).run(upstreamId);
return result.changes;
}
// ─── Phase 4b: 质量门实现 ────────────────────────────────
addQualityGate(gate: Omit<QualityGate, "id"> & { id?: string }): QualityGate {
const id = gate.id ?? `qg_${Date.now()}_${Math.random().toString(36).slice(2, 8)}`;
const fullGate: QualityGate = { ...gate, id };
db.prepare(
`INSERT INTO quality_gates (id, pipeline_id, gate_name, criteria, after_order, status, evaluator_id, result, evaluated_at, created_at)
VALUES (@id, @pipeline_id, @gate_name, @criteria, @after_order, @status, @evaluator_id, @result, @evaluated_at, @created_at)`
).run(fullGate);
return fullGate;
}
updateQualityGateStatus(gateId: string, status: GateStatus, evaluatorId: string, result?: string): void {
const now = Date.now();
db.prepare(
`UPDATE quality_gates SET status=?, evaluator_id=?, result=?, evaluated_at=? WHERE id=?`
).run(status, evaluatorId, result ?? null, now, gateId);
}
listGatesByPipeline(pipelineId: string): QualityGate[] {
return db.prepare(
`SELECT * FROM quality_gates WHERE pipeline_id=? ORDER BY after_order ASC`
).all(pipelineId) as QualityGate[];
}
}
// ─── ConsumedLogRepo ──────────────────────────────────────────
class SqliteConsumedLogRepo implements IConsumedLogRepo {
insert(entry: ConsumedEntry): void {
consumedStmt.insert.run(entry);
}
check(agentId: string, resource: string): ConsumedEntry | undefined {
return consumedStmt.check.get(agentId, resource) as ConsumedEntry | undefined;
}
listByAgent(agentId: string, limit = 50): ConsumedEntry[] {
return consumedStmt.listByAgent.all(agentId, limit) as ConsumedEntry[];
}
}
// ─── 单例导出 ──────────────────────────────────────────────────
export const messageRepo: IMessageRepo = new SqliteMessageRepo();
export const taskRepo: ITaskRepo = new SqliteTaskRepo();
export const consumedRepo: IConsumedLogRepo = new SqliteConsumedLogRepo();
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 |