agents-relay 1.0.158 → 1.0.160

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/cli.js CHANGED
@@ -7,8 +7,8 @@ import { dashboardManagedJobs, InMemoryStore, GitHubStore, resolvePullRequest, l
7
7
  import { githubAuthContext } from './github-auth.js';
8
8
  import { taskOverallTimeoutMs } from './types.js';
9
9
  import { Reconciler } from './reconciler.js';
10
- import { dashboardTaskAction, dashboardTaskSetState } from './dashboard-actions.js';
11
- export { dashboardTaskAction, dashboardTaskSetState } from './dashboard-actions.js';
10
+ import { dashboardJobSetState, dashboardTaskAction, dashboardTaskSetState } from './dashboard-actions.js';
11
+ export { dashboardJobSetState, dashboardTaskAction, dashboardTaskSetState } from './dashboard-actions.js';
12
12
  import { CodexAdapter, ChatGptAdapter } from './adapters.js';
13
13
  import { CodexThreadContinuation, CommandContinuation, WebhookContinuation } from './continuation.js';
14
14
  import { eventFor } from './events.js';
@@ -488,6 +488,7 @@ export async function createService(args) {
488
488
  start: dashboardJobAction(loaded.store, reconciler, 'start'),
489
489
  reconcile: dashboardJobAction(loaded.store, reconciler, 'reconcile'),
490
490
  cancel: dashboardJobAction(loaded.store, reconciler, 'cancel'),
491
+ jobSetState: dashboardJobSetState(loaded.store),
491
492
  taskRetry: dashboardTaskAction(loaded.store, reconciler, 'retry'),
492
493
  taskCancel: dashboardTaskAction(loaded.store, reconciler, 'cancel'),
493
494
  taskSetState: dashboardTaskSetState(loaded.store),
@@ -1,5 +1,5 @@
1
1
  import { extendRetryBudget, transitionTaskEvent } from './scheduler.js';
2
- import { applyTaskWorkflowDecision } from './workflow/runtime.js';
2
+ import { applyJobWorkflowDecision, applyTaskWorkflowDecision } from './workflow/runtime.js';
3
3
  export function dashboardTaskAction(store, reconciler, action) {
4
4
  return async (jobId, taskId) => {
5
5
  const job = await store.load(jobId);
@@ -26,6 +26,24 @@ export function dashboardTaskAction(store, reconciler, action) {
26
26
  return await reconciler.wake(job.id);
27
27
  };
28
28
  }
29
+ export function dashboardJobSetState(store) {
30
+ return async (jobId, state) => {
31
+ const job = await store.load(jobId);
32
+ if (job.state === state)
33
+ return job;
34
+ applyJobWorkflowDecision(job, {
35
+ accepted: true,
36
+ target: state,
37
+ guard: null,
38
+ actions: [],
39
+ mutations: [],
40
+ effects: [],
41
+ reason: 'accepted',
42
+ });
43
+ await store.saveJob(job);
44
+ return await store.load(job.id);
45
+ };
46
+ }
29
47
  export function dashboardTaskSetState(store) {
30
48
  return async (jobId, taskId, state) => {
31
49
  const job = await store.load(jobId);
@@ -324,13 +324,16 @@ function taskTagDisplay(job,task){const display=task.tag&&job?.tag?.values?.[tas
324
324
  function taskNode(task,children,depth){const li=element('li');const nested=children.get(task.id)||[];const line=element('div','task-line');if(nested.length){const key=(selectedJob||'')+':'+task.id;const collapsed=collapsedTasks.has(key);const fold=document.createElement('button');fold.type='button';fold.className='task-fold';fold.setAttribute('aria-expanded',String(!collapsed));fold.setAttribute('aria-label',(collapsed?'Expand ':'Collapse ')+task.id);fold.textContent=collapsed?'▸':'▾';line.append(fold);const ul=element('ul','tree');ul.hidden=collapsed;for(const child of nested)ul.append(taskNode(child,children,depth+1));fold.addEventListener('click',()=>{const next=!ul.hidden;ul.hidden=next;fold.setAttribute('aria-expanded',String(!next));fold.setAttribute('aria-label',(next?'Expand ':'Collapse ')+task.id);fold.textContent=next?'▸':'▾';if(next)collapsedTasks.add(key);else collapsedTasks.delete(key);});li.append(line,ul);}else li.append(line);const row=document.createElement('button');row.type='button';row.className='task-row';row.setAttribute('aria-current',String(task.id===selectedTaskId));row.title=task.id;row.addEventListener('click',()=>{const deselect=selectedTaskId===task.id;selectedTaskId=deselect?null:task.id;eventTaskId=deselect?null:task.id;rememberSelection();clearEvents();renderAll();void loadEventHistory();});const id=element('span','task-id',text(task.id));id.style.paddingLeft=Math.min(depth,4)*6+'px';row.append(id,element('span','badge '+value(task.state),text(task.state)));const tag=taskTagDisplay(selected(),task);if(tag){const badge=element('span','task-tag',text(tag));badge.title=(selected()?.tag?.description||'Task tag')+': '+task.tag;row.append(badge);}line.append(row);return li;}
325
325
  function buildTree(tasks){taskTree.replaceChildren();if(!Array.isArray(tasks)||!tasks.length){taskTree.append(element('li','empty',text('No tasks in durable state.')));return;}const work=tasks.filter(task=>task.kind!=='planner');if(!work.length){taskTree.append(element('li','empty',text('No work tasks in durable state.')));return;}const known=new Set(work.map(task=>task.id));const children=new Map();const roots=[];for(const task of work){if(!task.parentTaskId||!known.has(task.parentTaskId))roots.push(task);else{const siblings=children.get(task.parentTaskId)||[];siblings.push(task);children.set(task.parentTaskId,siblings);}}for(const task of roots)taskTree.append(taskNode(task,children,0));}
326
326
  function attemptBudget(task){const attempt=Number(task?.attempt)||0;const max=Number(task?.maxAttempts)||0;if(max>=3&&max%3===0){const effectiveMax=Math.max(max,Math.ceil(attempt/3)*3);return attempt+' / 3×'+(effectiveMax/3);}return attempt+' / '+max;}
327
+ const JOB_STATES=['DRAFT','ACTIVE','COMPLETED','CANCELLED'];
328
+ async function setJobState(job,state,select){if(!job||state===job.state)return;select.disabled=true;try{const response=await fetch('/api/job-state?id='+encodeURIComponent(job.id)+'&state='+encodeURIComponent(state),{method:'POST'});if(!response.ok)throw new Error(response.status+' '+response.statusText);const updated=await response.json();const index=jobs.findIndex(candidate=>candidate.id===job.id);if(index>=0)jobs[index]=updated;else jobs.push(updated);renderAll();void loadEventHistory();}catch(error){select.title='Set job lifecycle state: '+error.message;select.value=job.state;}finally{select.disabled=false;}}
329
+ function jobStateField(job){const box=element('div','detail-group');box.append(element('p','detail-label',text('Lifecycle state')));const select=document.createElement('select');select.className='job-filter job-state-select';select.setAttribute('aria-label','Job lifecycle state');for(const state of JOB_STATES)select.append(new Option(state,state));select.value=job.state;select.addEventListener('change',()=>void setJobState(job,select.value,select));box.append(select);return box;}
327
330
  const TASK_STATES=['DRAFT','QUEUED','RUNNING','BLOCKED','SUCCEEDED','FAILED','CANCELLED'];
328
331
  async function setTaskState(task,state,select){const job=selected();if(!job||state===task.state)return;select.disabled=true;try{const response=await fetch('/api/task-state?id='+encodeURIComponent(job.id)+'&task='+encodeURIComponent(task.id)+'&state='+encodeURIComponent(state),{method:'POST'});if(!response.ok)throw new Error(response.status+' '+response.statusText);const updated=await response.json();const index=jobs.findIndex(candidate=>candidate.id===job.id);if(index>=0)jobs[index]=updated;else jobs.push(updated);renderAll();void loadEventHistory();}catch(error){select.title='Set task state: '+error.message;select.value=task.state;}finally{select.disabled=false;}}
329
332
  function taskStateField(task){const box=element('div','detail-group');box.append(element('p','detail-label',text('State')));const select=document.createElement('select');select.className='job-filter task-state-select';select.setAttribute('aria-label','Task state');for(const state of TASK_STATES)select.append(new Option(state,state));select.value=task.state;select.addEventListener('change',()=>void setTaskState(task,select.value,select));box.append(select);return box;}
330
333
  function manualOverrideField(task){const rows=Array.isArray(task.manualStateOverrides)?task.manualStateOverrides:[];const last=rows.at(-1);return field('Manual override',last?last.from+' → '+last.to+' · '+relativeTime(last.at):'—');}
331
334
  function renderTaskDetail(task){taskDetailRoot.replaceChildren();if(!task){taskDetailRoot.append(element('p','empty',text('Select a task to inspect execution details.')));return;}const routing=task.routing||{};const groups=element('div','detail-groups');groups.append(field('Task',task.id),field('Kind',task.kind||'work'),taskStateField(task),manualOverrideField(task),field('Planning decision',task.planningDecisionId||'—'),field('Planning source',task.planningSource||'—'),field('Planning round',task.planningRound===undefined?'—':String(task.planningRound)),field('Project / agent',value(task.projectName)+' / '+value(task.agentName)),field('Parent task',task.parentTaskId||'root'),field('Adapter',task.adapter),field('Provider / model',value(routing.provider)+' / '+value(routing.model)),field('Reasoning / profile',value(routing.reasoning||routing.profile)),field('Attempt',attemptBudget(task)),field('Dependencies',Array.isArray(task.dependencies)&&task.dependencies.length?task.dependencies.join(', '):'none'),taskCommentField(task),chatField(task));const job=selected();if(task.reviewHeadSha&&job)groups.append(commitField('HEAD commit',job,task.reviewHeadSha));const graph=agentGraphField(task);if(graph)groups.append(graph);if(task.plannerResult)groups.append(field('Objective status',task.plannerResult.objective_status),markdownField('Assessment',task.plannerResult.assessment));if(task.error)groups.append(field('Error',task.error));taskDetailRoot.append(groups);}
332
335
  function reviewStatus(job){const review=job?.review;if(!review)return 'Not started';if(review.outcome==='approved'&&review.headSha===job.headSha)return 'MERGE READY · approved';const stale=Array.isArray(job?.tasks)&&job.tasks.some(task=>task.agentName==='reviewer'&&task.reviewHeadSha&&job.headSha&&task.reviewHeadSha!==job.headSha);if(stale)return 'STALE · awaiting review';return String(review.outcome||'pending').replace('_',' ').toUpperCase();}
333
- function renderSelectedJob(){jobDetail.replaceChildren();const job=selected();if(!job){selectedJobTitle.textContent='';jobDetail.append(element('div','empty',text('No durable jobs discovered.')));buildTree([]);renderTaskDetail(null);return;}selectedJobTitle.textContent=value(job.title||job.id);if(job.id!==selectedJob){selectedJob=job.id;selectedTaskId=null;eventTaskId=null;clearEvents();rememberSelection();}const summary=element('div','job-summary');summary.append(field('PR state',job.githubState||'UNKNOWN'),field('Lifecycle state',job.state),field('Runtime status',job.runtimeStatus||'IDLE'),field('Execution mode',job.executionMode||'fixed'),field('Cadence',cadence(job)),field('Review',reviewStatus(job)),commitField('Review head SHA',job,job.review?.headSha||job.headSha||''),prField(job),field('Repository',job.repository||'—'),field('Last activity',relativeTime(job.updatedAt)),field('Task states',compactCounts(job.tasks)));jobDetail.append(summary);const selectedTask=(job.tasks||[]).find(task=>task.id===selectedTaskId)||null;if(selectedTaskId&&!selectedTask){selectedTaskId=null;eventTaskId=null;clearEvents();rememberSelection();}buildTree(job.tasks||[]);renderTaskDetail(selectedTask);renderEventScope();}
336
+ function renderSelectedJob(){jobDetail.replaceChildren();const job=selected();if(!job){selectedJobTitle.textContent='';jobDetail.append(element('div','empty',text('No durable jobs discovered.')));buildTree([]);renderTaskDetail(null);return;}selectedJobTitle.textContent=value(job.title||job.id);if(job.id!==selectedJob){selectedJob=job.id;selectedTaskId=null;eventTaskId=null;clearEvents();rememberSelection();}const summary=element('div','job-summary');summary.append(field('PR state',job.githubState||'UNKNOWN'),jobStateField(job),field('Runtime status',job.runtimeStatus||'IDLE'),field('Execution mode',job.executionMode||'fixed'),field('Cadence',cadence(job)),field('Review',reviewStatus(job)),commitField('Review head SHA',job,job.review?.headSha||job.headSha||''),prField(job),field('Repository',job.repository||'—'),field('Last activity',relativeTime(job.updatedAt)),field('Task states',compactCounts(job.tasks)));jobDetail.append(summary);const selectedTask=(job.tasks||[]).find(task=>task.id===selectedTaskId)||null;if(selectedTaskId&&!selectedTask){selectedTaskId=null;eventTaskId=null;clearEvents();rememberSelection();}buildTree(job.tasks||[]);renderTaskDetail(selectedTask);renderEventScope();}
334
337
  function renderAll(){jobs.sort(compareJobs);syncFilters();renderJobList();renderSelectedJob();syncMobileJobsView();}
335
338
  function isCompleteUnmerged(job){return job.runtimeStatus==='MERGE_PENDING';}
336
339
  function isTerminalJob(job){return ['COMPLETED','CANCELLED'].includes(job.state)||['MERGED','CLOSED'].includes(job.githubState);}
package/dist/dashboard.js CHANGED
@@ -269,6 +269,31 @@ export function serveDashboard(store, port = 8787, bus, refreshMs = 300000, jobI
269
269
  }
270
270
  return;
271
271
  }
272
+ if (new URL(req.url ?? '/', 'http://127.0.0.1').pathname === '/api/job-state') {
273
+ const url = new URL(req.url ?? '/', 'http://127.0.0.1');
274
+ const requestedJobId = url.searchParams.get('id') ?? '';
275
+ const requestedState = url.searchParams.get('state') ?? '';
276
+ const validState = ['DRAFT', 'ACTIVE', 'COMPLETED', 'CANCELLED'].includes(requestedState);
277
+ if (req.method !== 'POST' || !/^[A-Za-z0-9_-]+$/.test(requestedJobId) || !validState || !resolvedActions?.jobSetState) {
278
+ res.writeHead(404);
279
+ res.end('Job state override unavailable');
280
+ return;
281
+ }
282
+ try {
283
+ const result = await resolvedActions.jobSetState(requestedJobId, requestedState);
284
+ if (!res.headersSent)
285
+ res.writeHead(200, { 'content-type': 'application/json' });
286
+ if (!res.writableEnded)
287
+ res.end(JSON.stringify(result));
288
+ }
289
+ catch (error) {
290
+ if (!res.headersSent)
291
+ res.writeHead(409);
292
+ if (!res.writableEnded)
293
+ res.end(error instanceof Error ? error.message : String(error));
294
+ }
295
+ return;
296
+ }
272
297
  if (new URL(req.url ?? '/', 'http://127.0.0.1').pathname === '/api/task-state') {
273
298
  const url = new URL(req.url ?? '/', 'http://127.0.0.1');
274
299
  const requestedJobId = url.searchParams.get('id') ?? '';
package/dist/relayd.js CHANGED
@@ -12,7 +12,7 @@ import { Reconciler } from './reconciler.js';
12
12
  import { dashboardManagedJobs, GitHubStore, InMemoryStore, loadManagedGitHubPullRequestSnapshotGraphql, withGitHubRequestContext } from './store.js';
13
13
  import { githubAuthContext } from './github-auth.js';
14
14
  import { dashboardJobAction, isEntrypoint, runtimePlanner, runtimeReviewer, startService, SERVICE_DEFAULTS } from './cli.js';
15
- import { dashboardTaskAction, dashboardTaskSetState } from './dashboard-actions.js';
15
+ import { dashboardJobSetState, dashboardTaskAction, dashboardTaskSetState } from './dashboard-actions.js';
16
16
  import { codexAndZaiUsageRegistry } from './usage.js';
17
17
  import { defaultWorkspaceRoot, discoverWorkspaceRepositories, resolveManagedProjectPath, resolveRepositoryProjectPath } from './workspace.js';
18
18
  import { deriveJobRuntimeStatus } from './runtime-status.js';
@@ -200,6 +200,15 @@ export async function runDaemon(argv) {
200
200
  }
201
201
  return await dashboardJobAction(store, makeReconciler(store, concurrency), action)(jobId);
202
202
  };
203
+ const workspaceJobSetState = async (jobId, state) => {
204
+ const overview = cachedOverviews.find(item => item.job.id === jobId);
205
+ if (!overview)
206
+ throw new Error(`Managed job ${jobId} not found`);
207
+ const store = new GitHubStore(client, overview.job.repository, overview.job.prNumber, trusted);
208
+ const result = await dashboardJobSetState(store)(jobId, state);
209
+ await refreshPullRequest(overview.job.repository, overview.job.prNumber, 'dashboard.job-state', { force: true });
210
+ return result;
211
+ };
203
212
  const workspaceTaskAction = (action) => async (jobId, taskId) => {
204
213
  const overview = cachedOverviews.find(item => item.job.id === jobId);
205
214
  if (!overview)
@@ -255,6 +264,7 @@ export async function runDaemon(argv) {
255
264
  start: workspaceJobAction('start'),
256
265
  reconcile: workspaceJobAction('reconcile'),
257
266
  cancel: workspaceJobAction('cancel'),
267
+ jobSetState: workspaceJobSetState,
258
268
  taskRetry: workspaceTaskAction('retry'),
259
269
  taskCancel: workspaceTaskAction('cancel'),
260
270
  taskSetState: workspaceTaskSetState,
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "agents-relay",
3
- "version": "1.0.158",
3
+ "version": "1.0.160",
4
4
  "description": "Durable async agent jobs coordinated through GitHub pull requests",
5
5
  "type": "module",
6
6
  "bin": {
@@ -52,9 +52,48 @@ Keep these concepts distinct: observation is not root cause; hypothesis is not r
52
52
 
53
53
  ### Job-defined Task tag
54
54
 
55
- Canonical `Task.state` is reserved for Relay lifecycle only. A Job may optionally declare one `tag` contract with a human-readable `description` explaining what the tag means and when workers should set it, plus an allowed `values` map from durable tag values to display strings. A Task stores the reported durable value in `tag`.
56
-
57
- A model-backed worker reports the applicable value in its canonical terminal event under `data.tag`, for example `{"tag":"pass"}`. Relay validates it against `Job.tag.values`, persists only valid values into the Task marker's `tag`, and does not reinterpret the tag as lifecycle state. The transport `data.tag` is not duplicated inside `result.data`. Jobs without a tag contract are unchanged.
55
+ `Task.state` is reserved for Relay lifecycle only: `DRAFT`, `QUEUED`, `RUNNING`, `BLOCKED`, `SUCCEEDED`, `FAILED`, and `CANCELLED`. A Job may additionally define one optional `tag` contract for a Job-specific result or classification that should be visible on each Task without changing its lifecycle state.
56
+
57
+ The Job owns the meaning of the tag. Its trusted Job marker declares a human-readable `description` explaining what the tag means and when workers should set it, plus the allowed durable values and their dashboard display strings:
58
+
59
+ ```json
60
+ {
61
+ "tag": {
62
+ "description": "Deterministic E2E outcome. Set pass when the fixed E2E exits 0; set fail when it still fails after the required retry.",
63
+ "values": {
64
+ "pass": "✅",
65
+ "fail": "❌"
66
+ }
67
+ }
68
+ }
69
+ ```
70
+
71
+ A worker does **not** edit the Task marker directly. When the Job defines a tag and the tag applies to the work just performed, the worker reports one allowed durable value in the canonical terminal event as `data.tag`:
72
+
73
+ ```json
74
+ {
75
+ "type": "task.completed",
76
+ "data": {
77
+ "tag": "pass",
78
+ "exit_code": 0
79
+ }
80
+ }
81
+ ```
82
+
83
+ Relay validates `data.tag` against `Job.tag.values`. If the value is allowed, Relay persists it into the trusted Task marker as the Task's `tag`; undeclared or invalid values are ignored. The transport `data.tag` is not duplicated into `result.data`. The resulting Task metadata may therefore contain both independent concepts at once:
84
+
85
+ ```json
86
+ {
87
+ "state": "SUCCEEDED",
88
+ "tag": "pass"
89
+ }
90
+ ```
91
+
92
+ The dashboard always keeps the normal lifecycle badge. When a valid Task tag is present, it additionally renders the display string from `Job.tag.values` after the lifecycle status. For the example above this appears as `SUCCEEDED ✅`.
93
+
94
+ A tag must describe something specific to the Job contract, not Relay lifecycle. Examples include an E2E result (`pass` / `fail`), publication result, validation classification, or another small Job-defined outcome. The Job description must tell the worker when to set the tag and what each allowed value means. Jobs without a `tag` declaration behave exactly as before.
95
+
96
+ For the ChatGPT Browser Worker guardian, for example, the Job defines `pass -> ✅` and `fail -> ❌`; each hourly E2E child Task reports `data.tag` after running the fixed deterministic E2E, while the Task lifecycle independently records whether Relay execution itself completed, failed, or became blocked.
58
97
 
59
98
 
60
99
  ### Optional Review Chair Tasks