agents-relay 2.0.182 → 2.0.184

package/README.md

@@ -249,6 +249,13 @@ Every model-backed task declares a durable output with `--output task-pr` or `--
 
 For Codex and ChatGPT, **correlated worker events are authoritative lifecycle inputs**, not a separate durable store. Event publication uses the canonical Neo `events-bus` surface (`events__publish` for sandboxed/hosted workers, or `NEO_EVENTS_EMIT` for direct local workers); Agents Relay subscribes, validates correlation, applies graph transitions, persists results in RelayStore, and projects them to GitHub. Relay publishes `task.process.launched` only after successful adapter launch; `task.started` is worker-owned and must be a new, correlated start acknowledgement. Codex publishes `task.process.exited` when its synchronous runtime exits, while ChatGPT publishes `task.process.async_exited` after its one-shot submission runtime exits; neither substitutes for `task.completed`, `task.failed`, `task.blocked`, or `task.cancelled`. The executing agent must publish exactly one correlated terminal event. Relay persists the matching durable state only from that terminal event. Model-backed tasks require an event bus and time out explicitly if no terminal event arrives.
 
+External execution uses the single global `webhelm` executable installed by the
+WebHelm skill. Before starting Relay, install the global skill and its launcher
+(`npx skills add lalalic/browser-workspace -g -y -s webhelm`, then
+`~/.agents/skills/webhelm/scripts/install-global-cli.sh`). Relay checks
+`webhelm platform profile chatgpt` before browser workspace initialization.
+Do not invoke the old `browser-workspace` executable or a separate `uvx` CLI.
+
 The ChatGPT adapter uses Relay's `agents/chatgpt-worker.agent.md` lifecycle contract. A new worker uses `webhelm platform run chatgpt new-thread --session-id SESSION_ID --config <file>` in a caller-owned Session. This is a regular persistent ChatGPT thread, which supports the connected Neo MCP in the verified account. Relay closes the Session after submission, but the ChatGPT thread remains in history. Existing-thread review and handoff continue through `new-turn`. File paths are included in the worker message for Neo MCP reading; they are not uploaded as browser attachments. Submission receipts retain `status: submitted`; the event bus owns task completion.
 
 ~~~sh
@@ -300,7 +307,7 @@ For every Job, unblock evidence becomes durable recovery context. Relay retries
 
 ### Browser Workspace product boundary
 
-Agents Relay owns the persistent `Relay` browser workspace as product infrastructure. At daemon startup it runs `webhelm status`, dynamically imports the returned absolute `admin_helper`, and calls `ensureWorkspace("Relay", 6)`. This bootstrap is best-effort: Browser Workspace discovery/helper failures are logged and skipped so they do not prevent `agents-relayd` from starting; ChatGPT browser work may remain unavailable until Browser Workspace recovers. Ordinary ChatGPT worker activity uses the public Browser Workspace one-shot `platform run ... --auto-session` command; orchestrator-owned self-healing flows may still use explicit session commands. Relay does not create or delete workspaces through CLI lifecycle commands.
+Agents Relay owns the persistent `Relay` browser workspace as product infrastructure. At daemon startup it runs `webhelm status`, dynamically imports the returned absolute `admin_helper`, and calls `ensureWorkspace("Relay", 6)`. This bootstrap is best-effort: Browser Workspace discovery/helper failures are logged and skipped so they do not prevent `agents-relayd` from starting; ChatGPT browser work may remain unavailable until Browser Workspace recovers. New ChatGPT workers use a Relay-owned browser Session and close it after verified submission. Existing-thread `new-turn` messages use the durable async queue directly, without creating a browser Session. Relay does not create or delete workspaces through CLI lifecycle commands.
 
 ## Shared publication reporting
 

package/dist/adapters.js

@@ -1,5 +1,6 @@
 import { access, readFile, realpath, writeFile, unlink } from 'node:fs/promises';
 import { join, relative, resolve, sep, basename } from 'node:path';
+import { homedir } from 'node:os';
 import { DEFAULT_PROMPT_REFERENCE_DIRECTORY } from './agent-prompt-builder.js';
 import { pathToFileURL } from 'node:url';
 import { createHash, randomUUID } from 'node:crypto';
@@ -168,7 +169,7 @@ export function parseDriverJson(output) {
     throw new Error('browser-worker driver returned no submitted receipt');
 }
 function browserWorkspaceCli(override = process.env.AGENTS_RELAY_WEBHELM_COMMAND) {
-    return override ? { command: override, prefix: [] } : { command: 'uvx', prefix: ['--from', 'browser-workspace', 'webhelm'] };
+    return { command: override || join(homedir(), '.local', 'bin', 'webhelm'), prefix: [] };
 }
 function browserWorkspaceCommand(cli, args, stdin, signal) {
     return rawBrowserWorkspaceCommand(cli, args, stdin, signal);
@@ -270,7 +271,11 @@ async function accessiblePromptReferences(prompt) {
     return readable;
 }
 export async function ensureChatGptBrowserWorkspace(command = process.env.AGENTS_RELAY_WEBHELM_COMMAND) {
-    const status = parseWorkspaceJson(await browserWorkspaceCommand(browserWorkspaceCli(command), ['status']));
+    const cli = browserWorkspaceCli(command);
+    const profile = parseWorkspaceJson(await browserWorkspaceCommand(cli, ['platform', 'profile', 'chatgpt']));
+    if (profile.platform !== 'chatgpt')
+        throw new Error('Global WebHelm CLI cannot load the ChatGPT platform skill');
+    const status = parseWorkspaceJson(await browserWorkspaceCommand(cli, ['status']));
     if (typeof status.admin_helper !== 'string' || !status.admin_helper.startsWith('/')) {
         throw new Error('webhelm status returned no absolute admin_helper path');
     }
@@ -281,12 +286,16 @@ export async function ensureChatGptBrowserWorkspace(command = process.env.AGENTS
     await admin.ensureWorkspace(CHATGPT_BROWSER_WORKSPACE, CHATGPT_BROWSER_WORKSPACE_POOL_SIZE);
 }
 async function runBrowserWorker(cli, configFile, signal, action = 'new-thread') {
+    // Durable async turns are queued without browser sessions by the WebHelm runtime.
+    if (action === 'new-turn') {
+        return browserWorkspaceCommand(cli, ['platform', 'run', 'chatgpt', 'new-turn', '--mode', 'async', '--config', configFile], undefined, signal);
+    }
     const session = parseWorkspaceJson(await browserWorkspaceCommand(cli, ['session', 'start', '--workspace', CHATGPT_BROWSER_WORKSPACE], undefined, signal));
     const sessionId = session.session_id;
     if (typeof sessionId !== 'string' || !sessionId)
         throw new Error('webhelm session start returned no session_id');
     try {
-        return await browserWorkspaceCommand(cli, ['platform', 'run', 'chatgpt', action, ...(action === 'new-turn' ? ['--mode', 'async'] : []), '--session-id', sessionId, '--config', configFile], undefined, signal);
+        return await browserWorkspaceCommand(cli, ['platform', 'run', 'chatgpt', action, '--session-id', sessionId, '--config', configFile], undefined, signal);
     }
     finally {
         await browserWorkspaceCommand(cli, ['session', 'stop', sessionId]).catch(() => { });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agents-relay",
-  "version": "2.0.182",
+  "version": "2.0.184",
   "description": "Durable async agent jobs coordinated through GitHub pull requests",
   "repository": {
     "type": "git",

package/skills/agents-relay/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: agents-relay
-description: Use when work needs durable, multi-agent orchestration across turns or processes: plan and execute Jobs/Tasks, coordinate dependencies and reviews, recover failures, and track authoritative lifecycle events. Also use for content-generation workflows (research, story, production, human review, publication) that need explicit agent roles and handoff boundaries. Not for simple one-shot tasks with no durable coordination.
+description: Use when work needs durable, multi-agent orchestration across turns or processes - plan and execute Jobs/Tasks, coordinate dependencies and reviews, recover failures, and track authoritative lifecycle events. Also use for content-generation workflows (research, story, production, human review, publication) that need explicit agent roles and handoff boundaries. Not for simple one-shot tasks with no durable coordination.
 requires:
   skills:
     - daemon-service-manage