agent-relay 6.2.3 → 6.2.5

package/README.md

@@ -1,33 +1,55 @@
-![Agent Relay](./readme-banner.png)
+<img src="./readme-banner.png" alt="Agent Relay" height="392">
 
+**Website:** [agentrelay.com](https://agentrelay.com) · **Docs:** [agentrelay.com/docs](https://agentrelay.com/docs)
 
+<a href="https://www.npmjs.com/package/@agent-relay/sdk"><img alt="npm" src="https://img.shields.io/npm/v/@agent-relay/sdk"></a>
+<a href="https://github.com/AgentWorkforce/relay/actions/workflows/test.yml"><img alt="Tests" src="https://img.shields.io/github/actions/workflow/status/AgentWorkforce/relay/test.yml?branch=main&label=tests"></a>
+<a href="./LICENSE"><img alt="License" src="https://img.shields.io/badge/license-Apache--2.0-blue.svg"></a>
 
-<div align="center">
+</div>
 
-  
+## Multi Agent Orchestration
 
-[![Featured on OSSCAR](https://osscar.dev/api/badge?slug=agentworkforce)](https://osscar.dev/org/agentworkforce)
+Enable your Claude Code, Codex, OpenCode agent spawn agent teams that can communicate and collaborate. Not subagents, but real agents who
+could spawn their own subagents. This allows for powerful AI cross-collaboration so you can get the best harnesses + models working
+together.
 
-Agent Relay is real-time communication infrastructure for agent-to-agent work. Spawn agents from code, give them shared channels, direct messages, threads, reactions, and presence, and let them coordinate in the same workspace.
+## Benefits Over Subagents
 
-It is not a framework or a harness. Your agents keep running however they already run. Agent Relay is the communication layer that helps them talk to each other and take action together.
+1. The agent orchestrating has full insight what the spawned agents are doing. It can read the logs and steer mid turn if needed
+2. Enables advanced swarm techniques as each agent can communicate with each other and coordinate to form agent teams for different types: review/fix loops, adversarial/debate pairs, fan-out -> pipeline -> gather, or lead + workers to name a few
+3. Diversity of thought and implementation. Codex implement, Claude review, Gemini do the final verification leads to better results as different models + harnesses excel in different things.
+4. Review happens as a conversation between the live reviewer and the live implementer, not as a report handed back to the parent after each one finishes.
+5. Audit trail exists outside the agent and outside the parent. With the [Agent Relay Observer](https://agentrelay.com/observer) you get full auditability into every single DM and group message sent by the agents.
 
-**Website:** [agentrelay.com](https://agentrelay.com) · **Docs:** [agentrelay.com/docs](https://agentrelay.com/docs)
+## Get Started
 
-  <a href="https://www.npmjs.com/package/@agent-relay/sdk"><img alt="npm" src="https://img.shields.io/npm/v/@agent-relay/sdk"></a>
-  <a href="https://github.com/AgentWorkforce/relay/actions/workflows/test.yml"><img alt="Tests" src="https://img.shields.io/github/actions/workflow/status/AgentWorkforce/relay/test.yml?branch=main&label=tests"></a>
-  <a href="./LICENSE"><img alt="License" src="https://img.shields.io/badge/license-Apache--2.0-blue.svg"></a>
-</div>
+1. Install the agent-relay cli
+
+```
+curl -fsSL https://raw.githubusercontent.com/AgentWorkforce/relay/main/install.sh | bash
+
+```
+
+2. Install the skill
+
+```
+npx skills add https://github.com/agentworkforce/skills --skill orchestrating-agent-relay
+```
+
+3. Tell your agent to use it
+
+```
+use the orchestrating-agent-relay skill to spawn a claude and codex agent and [YOUR_TASK]
+```
 
+For single, well-scoped, one-shot tasks, subagents still win. Agent relay's advantages compound when work is multi-step, multi-role, long-running or needs independent verification.
 
-## Why Agent Relay
+## SDK
 
-- **Built for real-time coordination**: channels, messages, inboxes, reactions, and presence for agents that need to collaborate.
-- **Works with terminal-native agents**: use Claude Code, Codex, Gemini CLI, OpenCode, and other supported runtimes without changing how they run.
-- **SDK-first**: spawn agents programmatically, route work, wait for readiness, and manage lifecycles from TypeScript or Python.
-- **Useful from both code and tools**: wire Relay into apps, scripts, plugins, and local workflows.
+Use the Agent Relay SDK to spawn and control agents programmatically.
 
-## Install
+### Install
 
 **TypeScript / Node.js**
 
@@ -45,7 +67,7 @@ pip install agent-relay-sdk
 
 See the [Python SDK](./packages/sdk-py) for Python usage and adapters.
 
-## Quick example
+### Quick example
 
 ```typescript
 import { AgentRelay, Models } from '@agent-relay/sdk';
@@ -82,53 +104,17 @@ await relay.shutdown();
 
 Want more than a toy example? Start with:
 
-- [Introduction](./docs/introduction.md)
-- [CLI on the Relay](./docs/cli-on-the-relay.md)
-- [Examples](./examples/README.md)
-- [TypeScript SDK README](./packages/sdk/README.md)
-- [Python SDK README](./packages/sdk-py/README.md)
+- [Introduction](https://agentrelay.com/docs/introduction)
+- [TypeScript SDK README](https://agentrelay.com/docs/typescript-sdk)
+- [Python SDK README](https://agentrelay.com/docs/python-sdk)
 
-## What you can build
+### What you can build
 
 - Multi-agent coding flows with shared channels and worker handoffs
 - Agent inboxes for status updates, blockers, and review loops
 - Tooling that lets existing agents communicate without rewriting their runtime
 - Local or remote coordination patterns where multiple agents need shared context
 
-## Claude Code plugin
-
-Use Agent Relay directly inside Claude Code, no SDK required. The plugin adds multi-agent coordination via slash commands or natural language.
-
-```text
-/plugin marketplace add Agentworkforce/skills
-/plugin install claude-relay-plugin
-```
-
-Once installed, you can coordinate teams of agents with built-in skills:
-
-```text
-> /relay-team Refactor the auth module, split the middleware, update tests, and update docs
-> /relay-fanout Run linting fixes across all packages in the monorepo
-> /relay-pipeline Analyze the API logs, generate a summary report, then draft an email
-```
-
-Or just describe what you want in plain language:
-
-```text
-> Use relay fan-out to lint all packages in parallel
-> Split the migration into three relay workers, one for the schema, one for the API, and one for the frontend
-```
-
-See [docs/plugin-claude-code.md](./docs/plugin-claude-code.md) and the [plugin README](https://github.com/AgentWorkforce/skills/tree/main/plugins/claude-relay-plugin) for more.
-
-## Agent Relay CLI
-
-Install the CLI with:
-
-```bash
-curl -fsSL https://raw.githubusercontent.com/AgentWorkforce/relay/main/install.sh | bash
-```
-
 Then use Agent Relay to bring agents into a shared workspace and route work between them.
 
 ## Supported agents and runtimes
@@ -142,7 +128,7 @@ Agent Relay is designed for terminal-native agents and SDK-driven workflows. Thi
 
 The broader SDK and workflow surface also includes additional integrations in the codebase. See the package docs for details.
 
-## Development
+### Development
 
 If you want to work on the repo itself:
 
@@ -154,7 +140,6 @@ npm test
 
 Useful references:
 
-- [ARCHITECTURE.md](./ARCHITECTURE.md)
 - [CHANGELOG.md](./CHANGELOG.md)
 - [GitHub Issues](https://github.com/AgentWorkforce/relay/issues)
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agent-relay",
-  "version": "6.2.3",
+  "version": "6.2.5",
   "description": "Real-time agent-to-agent communication system",
   "type": "module",
   "main": "dist/index.cjs",
@@ -10,7 +10,6 @@
     "install.sh",
     "scripts/postinstall.js",
     "scripts/build-cjs.mjs",
-    "relay-snippets",
     "LICENSE",
     "README.md"
   ],
@@ -133,14 +132,14 @@
   },
   "homepage": "https://github.com/AgentWorkforce/relay#readme",
   "dependencies": {
-    "@agent-relay/cloud": "6.2.3",
-    "@agent-relay/config": "6.2.3",
-    "@agent-relay/hooks": "6.2.3",
-    "@agent-relay/sdk": "6.2.3",
-    "@agent-relay/telemetry": "6.2.3",
-    "@agent-relay/trajectory": "6.2.3",
-    "@agent-relay/user-directory": "6.2.3",
-    "@agent-relay/utils": "6.2.3",
+    "@agent-relay/cloud": "6.2.5",
+    "@agent-relay/config": "6.2.5",
+    "@agent-relay/hooks": "6.2.5",
+    "@agent-relay/sdk": "6.2.5",
+    "@agent-relay/telemetry": "6.2.5",
+    "@agent-relay/trajectory": "6.2.5",
+    "@agent-relay/user-directory": "6.2.5",
+    "@agent-relay/utils": "6.2.5",
     "@aws-sdk/client-s3": "3.1020.0",
     "@modelcontextprotocol/sdk": "^1.0.0",
     "@relayauth/core": "^0.1.2",
@@ -167,7 +166,6 @@
     "smol-toml": "^1.6.0",
     "ssh2": "^1.17.0",
     "tar": "^7.5.10",
-    "uuid": "^10.0.0",
     "ws": "^8.18.3",
     "yaml": "^2.7.0",
     "zod": "^3.23.8",
@@ -182,7 +180,6 @@
     "@types/express": "^5.0.6",
     "@types/node": "^22.19.3",
     "@types/ssh2": "^1.15.5",
-    "@types/uuid": "^10.0.0",
     "@types/ws": "^8.18.1",
     "@typescript-eslint/eslint-plugin": "^8.18.2",
     "@typescript-eslint/parser": "^8.18.2",

package/relay-snippets/agent-policy-snippet.md

@@ -1,40 +0,0 @@
-# Agent Policy
-
-You are operating under organizational agent policies. These policies govern your interactions with other agents and tools.
-
-## Your Permissions
-
-Check the policy service for your specific permissions. If no explicit restrictions are defined, you have full permissions.
-
-## General Rules
-
-1. **Spawn Authorization**: Only spawn agents you are authorized to spawn. Check with Lead before spawning if unsure.
-
-2. **Message Routing**: Only message agents you are authorized to communicate with. Use proper channels.
-
-3. **Tool Usage**: Only use tools you are authorized to use. Read-only operations are generally safer.
-
-4. **Rate Limits**: Respect rate limits on messages. Don't spam other agents.
-
-## Restricted Agents
-
-Workers and non-lead agents typically have these restrictions:
-- Cannot spawn other agents without Lead approval
-- Can only message Lead, Coordinator, and their assigned peers
-- Limited to read-only tools unless explicitly granted write access
-
-## Lead Agents
-
-Lead agents typically have elevated permissions:
-- Can spawn Worker agents
-- Can message all agents
-- Can use all tools
-- Responsible for enforcing policy on spawned agents
-
-## Enforcement
-
-Policy violations are blocked at runtime. If your action is blocked, you'll receive a denial message explaining why. Do not attempt to circumvent policy restrictions.
-
-## Checking Your Policy
-
-To see your current policy, ask Lead or check the dashboard at `/api/policy/:workspaceId`.

package/relay-snippets/agent-relay-protocol.md

@@ -1,64 +0,0 @@
-# Agent Relay Protocol (Internal)
-
-Advanced features for session continuity and trajectory tracking.
-
-## Session Continuity
-
-Use `mcp__relaycast__message_dm_send` with a continuity message to save state for session recovery:
-
-```
-mcp__relaycast__message_dm_send(to: "system", text: "KIND: continuity\nACTION: save\n\nCurrent task: Implementing user authentication\nCompleted: User model, JWT utils\nIn progress: Login endpoint")
-```
-
-### When to Save
-
-- Before long-running operations (builds, tests)
-- When switching task areas
-- Every 15-20 minutes of active work
-- Before ending session
-
-## Work Trajectories
-
-Record your work as a trajectory for future agents.
-
-### Starting Work
-
-```bash
-trail start "Implement user authentication"
-trail start "Fix login bug" --task "agent-relay-123"
-```
-
-### Recording Decisions
-
-```bash
-trail decision "Chose JWT over sessions" --reasoning "Stateless scaling"
-trail decision "Used existing auth middleware"
-```
-
-### Completing Work
-
-```bash
-trail complete --summary "Added JWT auth" --confidence 0.85
-```
-
-Confidence: 0.9+ (high), 0.7-0.9 (good), 0.5-0.7 (some uncertainty), <0.5 (needs review)
-
-### Abandoning Work
-
-```bash
-trail abandon --reason "Blocked by missing credentials"
-```
-
-## Cross-Project Messaging
-
-In bridge mode, use `project:agent` format with `mcp__relaycast__message_dm_send`:
-
-```
-mcp__relaycast__message_dm_send(to: "frontend:Designer", text: "Please update the login UI.")
-```
-
-Special targets:
-
-- `project:lead` - Lead agent of that project
-- `project:*` - Broadcast to project
-- `*:*` - Broadcast to all projects

package/relay-snippets/agent-relay-snippet.md

@@ -1,117 +0,0 @@
-# Agent Relay
-
-Real-time agent-to-agent messaging via MCP tools.
-
-## MCP Tools
-
-All agent communication uses MCP tools provided by the Relaycast MCP server.
-Tool names use dot-notation: Claude uses `mcp__relaycast__<category>_<action>`, other CLIs use `relaycast.<category>.<action>`.
-
-| Tool                              | Description                           |
-| --------------------------------- | ------------------------------------- |
-| `message.dm.send(to, text)`       | Send a DM to an agent                 |
-| `message_post(channel, text)`     | Post a message to a channel           |
-| `message.inbox.check()`           | Check your inbox for new messages     |
-| `agent_list()`                    | List online agents                    |
-| `agent_add(name, cli, task)`      | Spawn a new worker agent              |
-| `agent_remove(name)`              | Release/stop a worker agent           |
-
-## Sending Messages
-
-### Direct Messages
-
-```
-mcp__relaycast__message_dm_send(to: "Bob", text: "Can you review my code changes?")
-```
-
-### Channel Messages
-
-```
-mcp__relaycast__message_post(channel: "general", text: "The API endpoints are ready")
-```
-
-## Spawning & Releasing Agents
-
-### Spawn a Worker
-
-```
-mcp__relaycast__agent_add(name: "WorkerName", cli: "claude", task: "Task description here")
-```
-
-### CLI Options
-
-| CLI Value   | Description                  |
-| ----------- | ---------------------------- |
-| `claude`    | Claude Code (Anthropic)      |
-| `codex`     | Codex CLI (OpenAI)           |
-| `gemini`    | Gemini CLI (Google)          |
-| `opencode`  | OpenCode CLI (multi-model)   |
-| `aider`     | Aider coding assistant       |
-| `goose`     | Goose AI assistant           |
-
-### Release a Worker
-
-```
-mcp__relaycast__agent_remove(name: "WorkerName")
-```
-
-## Receiving Messages
-
-Messages appear as:
-
-```
-Relay message from Alice [abc123]: Content here
-```
-
-Channel messages include `[#channel]`:
-
-```
-Relay message from Alice [abc123] [#general]: Hello!
-```
-
-Reply to the channel shown, not the sender.
-
-## When You Are Spawned
-
-If you were spawned by another agent:
-
-1. Your first message is your task from your spawner
-2. Use `mcp__relaycast__message_dm_send` to reply to your spawner
-3. Report status to your spawner (your lead), not broadcast
-
-```
-mcp__relaycast__message_dm_send(to: "Lead", text: "ACK: Starting on the task.")
-```
-
-## Protocol
-
-- **ACK** when you receive a task: `ACK: Brief description`
-- **DONE** when complete: `DONE: What was accomplished`
-- Send status to your **lead**, not broadcast
-
-## Agent Naming (Local vs Bridge)
-
-**Local communication** uses plain agent names. The `project:` prefix is **ONLY** for cross-project bridge mode.
-
-| Context                | Correct                                                     | Incorrect                                              |
-| ---------------------- | ----------------------------------------------------------- | ------------------------------------------------------ |
-| Local (same project)   | `mcp__relaycast__message_dm_send(to: "Lead", ...)`          | `mcp__relaycast__message_dm_send(to: "project:lead", ...)`     |
-| Bridge (cross-project) | `mcp__relaycast__message_dm_send(to: "frontend:Designer", ...)` | N/A                                                    |
-
-## Multi-Workspace
-
-When connected to multiple workspaces, messages include workspace context:
-
-```
-Relay message from Alice [my-team / abc123]: Hello!
-```
-
-- Messages are scoped to the originating workspace
-- Reply within the same workspace context shown in the message header
-
-## Checking Status
-
-```
-mcp__relaycast__agent_list()          # List online agents
-mcp__relaycast__message_inbox_check() # Check for unread messages
-```