@yemi33/minions 0.1.1771 → 0.1.1773

package/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 0.1.1773 (2026-05-07)
+
+### Features
+- fix doc chat sticky scroll (#2176)
+- fix live output route and CLI docs (#2172)
+- make API project resolution strict (#2169)
+- fix copilot cc resume context (#2166)
+
 ## 0.1.1771 (2026-05-07)
 
 ### Features

package/README.md

@@ -127,21 +127,31 @@ minions work "Explore the codebase and document the architecture"
 | `minions init` | Bootstrap `~/.minions/` with default agents and config |
 | `minions update` | Update to latest version (npm update + apply) |
 | `minions version` | Show installed vs package version |
+| `minions doctor` | Check prerequisites and runtime health |
 | `minions scan [dir] [depth]` | Scan for git repos and multi-select to add (default: ~, depth 3) |
 | `minions add <dir>` | Link a single project (auto-detects settings from git, prompts to confirm) |
 | `minions remove <dir-or-name> [--keep-data \| --purge --force]` | Unlink a project: cancels pending work items, drains dispatch + kills active agents, cleans worktrees, disables linked schedules, archives `projects/<name>/` to `projects/.archived/<name>-YYYYMMDD/`. Use `--keep-data` to leave the data dir in place, or `--purge --force` to delete it. |
 | `minions list` | List all linked projects with descriptions |
+| `minions restart` | Start engine and dashboard together (recommended after reboot) |
 | `minions start` | Start engine daemon (ticks every 60s, auto-syncs MCP servers) |
 | `minions stop` | Stop the engine |
 | `minions status` | Show agents, projects, dispatch queue, quality metrics |
 | `minions pause` / `resume` | Pause/resume dispatching |
 | `minions dispatch` | Force a dispatch cycle |
 | `minions discover` | Dry-run work discovery |
+| `minions queue` | Show dispatch queue (pending, active, completed) |
+| `minions sources` | Show work source status per project |
 | `minions work <title> [opts]` | Add to central work queue |
 | `minions spawn <agent> <prompt>` | Manually spawn an agent |
 | `minions plan <file\|text> [proj]` | Run a plan |
+| `minions kill` | Kill all active agents and reset their dispatches to pending |
+| `minions complete <dispatch-id>` | Manually mark a dispatch as completed |
+| `minions config set-cli <R> [--model M]` | Persist the default runtime/model without starting the engine |
+| `minions mcp-sync` | Sync MCP servers from `~/.claude.json` |
 | `minions cleanup` | Run cleanup manually (temp files, worktrees, zombies) |
 | `minions dash` | Open dashboard (starts if not already running, port 7331) |
+| `minions nuke --confirm` | Factory reset runtime state and reset config to defaults |
+| `minions uninstall --confirm` | Remove Minions state and uninstall the npm package |
 
 You can also run scripts directly: `node ~/.minions/engine.js start`, `node ~/.minions/dashboard.js`, etc.
 

package/bin/minions.js

@@ -8,18 +8,29 @@
  *   minions add <project-dir>        Link a project (interactive)
  *   minions remove <project-dir>     Unlink a project
  *   minions list                     List linked projects
+ *   minions update                   Update to latest version
+ *   minions version                  Show installed and package versions
+ *   minions doctor                   Check prerequisites and runtime health
+ *   minions restart                  Start engine + dashboard
  *   minions start                    Start the engine
  *   minions stop                     Stop the engine
  *   minions status                   Show engine status
  *   minions pause / resume           Pause/resume dispatching
+ *   minions queue                    Show dispatch queue
+ *   minions sources                  Show work source status
  *   minions dash                     Start the dashboard
  *   minions work <title> [opts-json] Add a work item
  *   minions spawn <agent> <prompt>   Manually spawn an agent
  *   minions dispatch                 Force a dispatch cycle
  *   minions discover                 Dry-run work discovery
  *   minions cleanup                  Run cleanup manually
+ *   minions kill                     Kill active agents and reset to pending
+ *   minions complete <dispatch-id>   Mark a dispatch completed
+ *   minions config set-cli <R> [--model M]  Persist default runtime/model
  *   minions plan <file|text> [proj]  Run a plan
- *   minions version                  Show installed and package versions
+ *   minions mcp-sync                 Sync MCP servers from ~/.claude.json
+ *   minions nuke --confirm           Factory reset runtime state/config
+ *   minions uninstall --confirm      Remove Minions and uninstall package
  */
 
 const fs = require('fs');
@@ -642,6 +653,9 @@ if (!cmd || cmd === 'help' || cmd === '--help' || cmd === '-h') {
     minions plan <file|text> [proj]  Run a plan
     minions kill                     Kill all active agents and reset to pending
     minions complete <dispatch-id>   Manually mark a dispatch as completed
+    minions config set-cli <R> [--model M]
+                                      Persist default runtime/model without starting
+    minions mcp-sync                 Sync MCP servers from ~/.claude.json
     minions cleanup                  Clean temp files, worktrees, zombies
     minions nuke --confirm           Factory reset (delete state, reset config to defaults)
     minions uninstall --confirm      Remove everything + uninstall npm package

package/dashboard/js/modal-qa.js

@@ -30,6 +30,32 @@ const QA_QUEUE_CAP = 10; // max queued messages
 const QA_STREAM_STALL_MS = 6 * 60 * 1000; // allow the full doc-chat timeout before treating heartbeat-only streams as stalled
 let _qaSessionKey = ''; // key for current conversation (title or filePath)
 
+const QA_STICKY_BOTTOM_PX = 80;
+let _qaThreadShouldFollow = true;
+
+function _qaIsNearThreadBottom(thread) {
+  if (!thread) return true;
+  return thread.scrollHeight - thread.scrollTop - thread.clientHeight <= QA_STICKY_BOTTOM_PX;
+}
+
+function _qaShouldFollowThread(thread) {
+  return _qaThreadShouldFollow && _qaIsNearThreadBottom(thread);
+}
+
+function _qaSetThreadFollowFromScroll(thread) {
+  _qaThreadShouldFollow = _qaIsNearThreadBottom(thread);
+}
+
+function _qaScrollThreadToBottom(thread) {
+  if (!thread) return;
+  thread.scrollTop = thread.scrollHeight;
+  _qaThreadShouldFollow = true;
+}
+
+function _qaMaybeScrollThreadToBottom(thread, shouldFollow) {
+  if (shouldFollow) _qaScrollThreadToBottom(thread);
+}
+
 // Insert html at the bottom of the thread but above any pending "Queued: ..."
 // strips, so queued messages always remain visually below the active progress
 // UX / answer / errors.
@@ -47,7 +73,7 @@ function _renderQaUserMessage(thread, message, selection) {
   }
   qHtml += '</div>';
   _qaInsertBeforeQueued(thread, qHtml);
-  thread.scrollTop = thread.scrollHeight;
+  _qaScrollThreadToBottom(thread);
   _showThreadWrap();
 }
 const _qaSessions = new Map(); // persist conversations across modal open/close {key → {history, threadHtml}}
@@ -91,6 +117,10 @@ function _qaThreadEl() {
   return document.getElementById('modal-qa-thread');
 }
 
+document.addEventListener('scroll', function(e) {
+  if (e.target && e.target.id === 'modal-qa-thread') _qaSetThreadFollowFromScroll(e.target);
+}, true);
+
 function _qaThreadHtml() {
   return (_qaThreadEl() || {}).innerHTML || '';
 }
@@ -308,8 +338,9 @@ function _qaMutateThreadHtml(key, mutate) {
     const wasCollapsed = _qaIsThreadCollapsed();
     const thread = _qaThreadEl();
     if (thread) {
+      const shouldFollow = _qaShouldFollowThread(thread);
       thread.innerHTML = html;
-      thread.scrollTop = thread.scrollHeight;
+      _qaMaybeScrollThreadToBottom(thread, shouldFollow);
     }
     if (wasCollapsed) _setQaThreadCollapsed(true);
     else _showThreadWrap();
@@ -386,7 +417,7 @@ function _initQaSession() {
     _showThreadWrap();
     requestAnimationFrame(function() {
       var thread = document.getElementById('modal-qa-thread');
-      if (thread) thread.scrollTop = thread.scrollHeight;
+      _qaScrollThreadToBottom(thread);
     });
     if (_qaQueue.length > 0 && !_qaProcessing) {
       setTimeout(_qaResumeQueuedMessages, 0);
@@ -394,6 +425,7 @@ function _initQaSession() {
   } else {
     _qaHistory = [];
     document.getElementById('modal-qa-thread').innerHTML = '';
+    _qaThreadShouldFollow = true;
     var wrap = document.getElementById('modal-qa-thread-wrap');
     var expandBar = document.getElementById('qa-expand-bar');
     if (wrap) wrap.style.display = 'none';
@@ -410,6 +442,7 @@ function clearQaConversation() {
   _qaProcessing = false;
   _qaAbortController = null;
   document.getElementById('modal-qa-thread').innerHTML = '';
+  _qaThreadShouldFollow = true;
   var wrap = document.getElementById('modal-qa-thread-wrap');
   var expandBar = document.getElementById('qa-expand-bar');
   if (wrap) wrap.style.display = 'none';
@@ -455,7 +488,7 @@ function modalSend() {
     }
     _qaQueue.push({ message, selection });
     thread.insertAdjacentHTML('beforeend', _qaBuildQueuedHtml(message));
-    thread.scrollTop = thread.scrollHeight;
+    _qaScrollThreadToBottom(thread);
     _showThreadWrap();
     _qaSaveActiveSessionState();
     return;

package/dashboard.js

@@ -10,6 +10,7 @@ const zlib = require('zlib');
 const fs = require('fs');
 const path = require('path');
 const llm = require('./engine/llm');
+const { resolveRuntime } = require('./engine/runtimes');
 
 // Dashboard version stamp — captured at module load so it reflects the code actually running
 const _dashboardVersion = {
@@ -280,12 +281,23 @@ function createWorkItemWithDedup(wiPath, item, options = {}) {
   return result || { created: false, item: null };
 }
 
+function formatUnknownProjectError(projectName, projects = []) {
+  const known = projects.map(p => p.name).filter(Boolean).join(', ') || '(none configured)';
+  return `Project "${projectName}" not found. Known projects: ${known}`;
+}
+
+function findProjectByName(projects, projectName) {
+  const name = String(projectName || '').trim().toLowerCase();
+  if (!name) return null;
+  return projects.find(p => p.name?.toLowerCase() === name) || null;
+}
+
 function resolveWorkItemsCreateTarget(projectName, projects = PROJECTS) {
   const project = String(projectName || '').trim();
   let targetProject = null;
   if (project) {
-    targetProject = projects.find(p => p.name === project) || (projects.length > 0 ? projects[0] : null);
-    if (!targetProject) return { error: 'No projects configured' };
+    targetProject = findProjectByName(projects, project);
+    if (!targetProject) return { error: formatUnknownProjectError(project, projects) };
   } else if (projects.length === 1) {
     targetProject = projects[0];
   }
@@ -325,7 +337,13 @@ function linkPullRequestForTracking({ url, title, project: projectName, autoObse
     throw err;
   }
   const projects = shared.getProjects(config);
-  const targetProject = projectName ? projects.find(p => p.name?.toLowerCase() === projectName.toLowerCase()) : (projects[0] || null);
+  const explicitProjectName = String(projectName || '').trim();
+  const targetProject = explicitProjectName ? findProjectByName(projects, explicitProjectName) : (projects[0] || null);
+  if (explicitProjectName && !targetProject) {
+    const err = new Error(formatUnknownProjectError(explicitProjectName, projects));
+    err.statusCode = 400;
+    throw err;
+  }
   const prPath = targetProject ? shared.projectPrPath(targetProject) : path.join(MINIONS_DIR, 'pull-requests.json');
 
   const prNumMatch = url.match(/\/pull\/(\d+)|pullrequest\/(\d+)/);
@@ -1262,9 +1280,14 @@ function _readCcTabSessions({ prune = true } = {}) {
 const CC_CARRYOVER_MAX_TURNS = 20;
 const CC_CARRYOVER_PER_MSG_CHARS = 2000;
 
-function _buildTranscriptCarryover(transcript, { previousRuntime } = {}) {
+function _buildTranscriptCarryover(transcript, { previousRuntime, currentMessage } = {}) {
   if (!Array.isArray(transcript) || transcript.length === 0) return '';
-  const filtered = transcript.filter(m => m && (m.role === 'user' || m.role === 'assistant') && typeof m.text === 'string' && m.text.trim());
+  let filtered = transcript.filter(m => m && (m.role === 'user' || m.role === 'assistant') && typeof m.text === 'string' && m.text.trim());
+  const current = typeof currentMessage === 'string' ? currentMessage.trim() : '';
+  if (current && filtered.length > 0) {
+    const last = filtered[filtered.length - 1];
+    if (last.role === 'user' && last.text.trim() === current) filtered = filtered.slice(0, -1);
+  }
   if (filtered.length === 0) return '';
   const recent = filtered.slice(-CC_CARRYOVER_MAX_TURNS);
   const truncated = filtered.length > recent.length;
@@ -1281,6 +1304,19 @@ function _buildTranscriptCarryover(transcript, { previousRuntime } = {}) {
   return `${header}\n\n${truncationNote}${lines.join('\n\n')}\n\n--- Current message follows ---`;
 }
 
+function _ccRuntimeNeedsResumeCarryover(runtimeName) {
+  try {
+    const runtime = resolveRuntime(runtimeName);
+    return !!runtime?.capabilities?.resumePromptCarryover;
+  } catch {
+    return false;
+  }
+}
+
+function _joinCcPromptParts(...parts) {
+  return parts.filter(Boolean).join('\n\n---\n\n');
+}
+
 // Load persisted CC session on startup. CC chat sessions are non-expiring;
 // only restore-time validity checks here are sessionId presence (anything
 // else would auto-expire the user's chat without their consent).
@@ -2415,7 +2451,7 @@ async function _preflightModelCheck({ runtime: cliOverride, model: modelOverride
  * @param {number} opts.maxTurns - Max tool-use turns
  * @param {string} opts.allowedTools - Comma-separated tool list
  */
-async function ccCall(message, { store = 'cc', sessionKey, extraContext, label = 'command-center', timeout = CC_CALL_TIMEOUT_MS, maxTurns, allowedTools = 'Bash,Read,Write,Edit,Glob,Grep,WebFetch,WebSearch', skipStatePreamble = false, model, onAbortReady, systemPrompt = CC_STATIC_SYSTEM_PROMPT } = {}) {
+async function ccCall(message, { store = 'cc', sessionKey, extraContext, label = 'command-center', timeout = CC_CALL_TIMEOUT_MS, maxTurns, allowedTools = 'Bash,Read,Write,Edit,Glob,Grep,WebFetch,WebSearch', skipStatePreamble = false, model, onAbortReady, systemPrompt = CC_STATIC_SYSTEM_PROMPT, transcript } = {}) {
   if (!maxTurns) maxTurns = CONFIG.engine?.ccMaxTurns || shared.ENGINE_DEFAULTS.ccMaxTurns;
   if (!model) model = CONFIG.engine?.ccModel || shared.ENGINE_DEFAULTS.ccModel;
   const ccEffort = CONFIG.engine?.ccEffort || shared.ENGINE_DEFAULTS.ccEffort;
@@ -2428,10 +2464,15 @@ async function ccCall(message, { store = 'cc', sessionKey, extraContext, label =
 
   const existing = resolveSession(store, sessionKey);
   let sessionId = existing ? existing.sessionId : null;
+  const resumeNeedsCarryover = !!sessionId && _ccRuntimeNeedsResumeCarryover(shared.resolveCcCli(CONFIG.engine));
 
-  function buildPrompt({ includePreamble = true } = {}) {
+  function buildPrompt({ includePreamble = true, includeCarryover = false } = {}) {
     const parts = (!skipStatePreamble && includePreamble) ? [`## Current Minions State (${new Date().toISOString().slice(0, 16)})\n\n${buildCCStatePreamble()}`] : [];
     if (extraContext) parts.push(extraContext);
+    if (includeCarryover) {
+      const carryover = _buildTranscriptCarryover(transcript, { currentMessage: message });
+      if (carryover) parts.push(carryover);
+    }
     parts.push(message);
     return parts.join('\n\n---\n\n');
   }
@@ -2440,7 +2481,7 @@ async function ccCall(message, { store = 'cc', sessionKey, extraContext, label =
 
   // Attempt 1: resume existing session — skip preamble (session already has context)
   if (sessionId && maxTurns > 1) {
-    const p1 = llm.callLLM(buildPrompt({ includePreamble: false }), '', {
+    const p1 = llm.callLLM(buildPrompt({ includePreamble: false, includeCarryover: resumeNeedsCarryover }), '', {
       timeout, label, model, maxTurns, allowedTools, sessionId, effort: ccEffort, direct: true,
       engineConfig: CONFIG.engine,
     });
@@ -2477,7 +2518,7 @@ async function ccCall(message, { store = 'cc', sessionKey, extraContext, label =
   }
 
   // Attempt 2: fresh session (include preamble for full context)
-  const freshPrompt = buildPrompt();
+  const freshPrompt = buildPrompt({ includeCarryover: resumeNeedsCarryover });
   const p2 = llm.callLLM(freshPrompt, systemPrompt, {
     timeout, label, model, maxTurns, allowedTools, effort: ccEffort, direct: true,
     engineConfig: CONFIG.engine,
@@ -2511,7 +2552,7 @@ async function ccCall(message, { store = 'cc', sessionKey, extraContext, label =
   return result;
 }
 
-async function ccCallStreaming(message, { store = 'cc', sessionKey, extraContext, label = 'command-center', timeout = CC_CALL_TIMEOUT_MS, maxTurns, allowedTools = 'Bash,Read,Write,Edit,Glob,Grep,WebFetch,WebSearch', skipStatePreamble = false, model, onAbortReady, onChunk, onToolUse, onRetry, systemPrompt = CC_STATIC_SYSTEM_PROMPT } = {}) {
+async function ccCallStreaming(message, { store = 'cc', sessionKey, extraContext, label = 'command-center', timeout = CC_CALL_TIMEOUT_MS, maxTurns, allowedTools = 'Bash,Read,Write,Edit,Glob,Grep,WebFetch,WebSearch', skipStatePreamble = false, model, onAbortReady, onChunk, onToolUse, onRetry, systemPrompt = CC_STATIC_SYSTEM_PROMPT, transcript } = {}) {
   if (!maxTurns) maxTurns = CONFIG.engine?.ccMaxTurns || shared.ENGINE_DEFAULTS.ccMaxTurns;
   if (!model) model = CONFIG.engine?.ccModel || shared.ENGINE_DEFAULTS.ccModel;
   const ccEffort = CONFIG.engine?.ccEffort || shared.ENGINE_DEFAULTS.ccEffort;
@@ -2524,10 +2565,15 @@ async function ccCallStreaming(message, { store = 'cc', sessionKey, extraContext
 
   const existing = resolveSession(store, sessionKey);
   let sessionId = existing ? existing.sessionId : null;
+  const resumeNeedsCarryover = !!sessionId && _ccRuntimeNeedsResumeCarryover(shared.resolveCcCli(CONFIG.engine));
 
-  function buildPrompt({ includePreamble = true } = {}) {
+  function buildPrompt({ includePreamble = true, includeCarryover = false } = {}) {
     const parts = (!skipStatePreamble && includePreamble) ? [`## Current Minions State (${new Date().toISOString().slice(0, 16)})\n\n${buildCCStatePreamble()}`] : [];
     if (extraContext) parts.push(extraContext);
+    if (includeCarryover) {
+      const carryover = _buildTranscriptCarryover(transcript, { currentMessage: message });
+      if (carryover) parts.push(carryover);
+    }
     parts.push(message);
     return parts.join('\n\n---\n\n');
   }
@@ -2535,7 +2581,7 @@ async function ccCallStreaming(message, { store = 'cc', sessionKey, extraContext
   let result;
 
   if (sessionId && maxTurns > 1) {
-    const p1 = llm.callLLMStreaming(buildPrompt({ includePreamble: false }), '', {
+    const p1 = llm.callLLMStreaming(buildPrompt({ includePreamble: false, includeCarryover: resumeNeedsCarryover }), '', {
       timeout, label, model, maxTurns, allowedTools, sessionId, effort: ccEffort, direct: true,
       engineConfig: CONFIG.engine,
       onChunk,
@@ -2572,7 +2618,7 @@ async function ccCallStreaming(message, { store = 'cc', sessionKey, extraContext
   }
 
   if (onRetry) onRetry(2);
-  const freshPrompt = buildPrompt();
+  const freshPrompt = buildPrompt({ includeCarryover: resumeNeedsCarryover });
   const p2 = llm.callLLMStreaming(freshPrompt, systemPrompt, {
     timeout, label, model, maxTurns, allowedTools, effort: ccEffort, direct: true,
     engineConfig: CONFIG.engine,
@@ -5754,7 +5800,7 @@ What would you like to discuss or change? When you're happy, say "approve" and I
         }
         const wasResume = !!(body.sessionId && body.sessionId === ccSession.sessionId && ccSessionValid());
 
-        const result = await ccCall(body.message, { store: 'cc' });
+        const result = await ccCall(body.message, { store: 'cc', transcript: body.transcript });
 
         // Non-zero exit with text = max_turns or partial success — still usable
         if (!result.text) {
@@ -5989,13 +6035,21 @@ What would you like to discuss or change? When you're happy, say "approve" and I
             sessionReset = true;
             sessionResetReason = 'runtimeChanged';
             previousRuntime = tabEntry.runtime;
+          } else if (tabEntry.sessionId && tabEntry.sessionId !== tabSessionId) {
+            tabSessionId = tabEntry.sessionId;
           }
         }
         const wasResume = !!tabSessionId;
         const sessionId = tabSessionId || null;
+        const resumeNeedsCarryover = wasResume && _ccRuntimeNeedsResumeCarryover(currentRuntime);
         const preamble = wasResume ? '' : buildCCStatePreamble();
-        const carryover = sessionReset ? _buildTranscriptCarryover(body.transcript, { previousRuntime }) : '';
-        const prompt = (preamble ? preamble + '\n\n---\n\n' : '') + (carryover ? carryover + '\n\n---\n\n' : '') + body.message;
+        const carryover = (sessionReset || resumeNeedsCarryover)
+          ? _buildTranscriptCarryover(body.transcript, {
+            previousRuntime: sessionReset ? previousRuntime : null,
+            currentMessage: body.message,
+          })
+          : '';
+        const prompt = _joinCcPromptParts(preamble, carryover, body.message);
 
         const { trackEngineUsage: trackUsage } = require('./engine/llm');
         const streamModel = CONFIG.engine?.ccModel || shared.ENGINE_DEFAULTS.ccModel;
@@ -6023,7 +6077,8 @@ What would you like to discuss or change? When you're happy, say "approve" and I
           // Resume failed (stale/expired session) — auto-retry as fresh session (skip if client already disconnected)
           console.log(`[CC-stream] Resume failed (code=${result.code}) — retrying fresh`);
           const freshPreamble = buildCCStatePreamble();
-          const freshPrompt = (freshPreamble ? freshPreamble + '\n\n---\n\n' : '') + body.message;
+          const freshCarryover = _buildTranscriptCarryover(body.transcript, { currentMessage: body.message });
+          const freshPrompt = _joinCcPromptParts(freshPreamble, freshCarryover, body.message);
           toolUses = []; // discard stale metadata from the failed resume attempt
           const retryPromise = _invokeCcStream({
             prompt: freshPrompt, sessionId: undefined, liveState, toolUses,
@@ -7083,6 +7138,13 @@ What would you like to discuss or change? When you're happy, say "approve" and I
       if (!url) return jsonReply(res, 400, { error: 'url required' });
 
       reloadConfig();
+      const explicitProjectName = String(body.project || '').trim();
+      if (explicitProjectName) {
+        const projects = shared.getProjects(CONFIG);
+        if (!findProjectByName(projects, explicitProjectName)) {
+          return jsonReply(res, 400, { error: formatUnknownProjectError(explicitProjectName, projects) }, req);
+        }
+      }
       const adoTarget = parseAdoPrMetadataTarget(url);
       let initialPrData = null;
       if (adoTarget) {
@@ -7092,7 +7154,13 @@ What would you like to discuss or change? When you're happy, say "approve" and I
           shared.log('warn', `ADO PR link metadata fetch failed for ${url}: ${e.message}`);
         }
       }
-      const { id: prId, prPath, prNum, created, linked } = linkPullRequestForTracking(body, CONFIG, { metadata: initialPrData });
+      let linkResult;
+      try {
+        linkResult = linkPullRequestForTracking(body, CONFIG, { metadata: initialPrData });
+      } catch (e) {
+        return jsonReply(res, e.statusCode || 400, { error: e.message }, req);
+      }
+      const { id: prId, prPath, prNum, created, linked } = linkResult;
       invalidateStatusCache();
       jsonReply(res, 200, { ok: true, id: prId, created, linked });
 
@@ -7230,6 +7298,7 @@ What would you like to discuss or change? When you're happy, say "approve" and I
     { method: 'POST', path: /^\/api\/agent\/([\w-]+)\/kill$/, desc: 'Kill a running agent: stop process, clear dispatch, reset work items to pending', handler: handleAgentKill },
     { method: 'GET', path: /^\/api\/agent\/([\w-]+)\/live-stream(?:\?.*)?$/, desc: 'SSE real-time live output streaming', handler: handleAgentLiveStream },
     { method: 'GET', path: /^\/api\/agent\/([\w-]+)\/live(?:\?.*)?$/, desc: 'Tail live output for a working agent', params: 'tail? (bytes, default 8192)', handler: handleAgentLive },
+    { method: 'GET', path: /^\/api\/agent\/([\w-]+)\/live-output(?:\?.*)?$/, desc: 'Tail live output for a working agent (alias for /live)', params: 'tail? (bytes, default 8192)', handler: handleAgentLive },
     { method: 'GET', path: /^\/api\/agent\/([\w-]+)\/output(?:\?.*)?$/, desc: 'Fetch final output.log for an agent', handler: handleAgentOutput },
     { method: 'GET', path: /^\/api\/agent\/([\w-]+)$/, desc: 'Get detailed agent info', handler: handleAgentDetail },
     { method: 'GET', path: /^\/api\/dispatch\/([\w.-]+)\/completion-report$/, desc: 'Read structured completion report for a dispatch', handler: (req, res, match) => {
@@ -7695,6 +7764,9 @@ module.exports = {
   _createPipelineFromAction: createPipelineFromAction,
   executeCCActions,
   buildCCStatePreamble,
+  _buildTranscriptCarryover,
+  _ccRuntimeNeedsResumeCarryover,
+  _joinCcPromptParts,
   _captureApiRoutesMeta,
   _formatCcApiRoutesIndex,
   _formatCcCliCommandsIndex,

package/docs/copilot-cli-schema.md

@@ -17,6 +17,7 @@
 | `capabilities.modelDiscovery` | **`true`** | `GET https://api.githubcopilot.com/models` with a `gh auth token` Bearer returns HTTP 200 + a 24-model JSON catalog. |
 | `capabilities.streaming` | **`true`** | `--stream on` (default) emits `assistant.message_delta` events incrementally; `--stream off` suppresses deltas but the final `assistant.message` always arrives. |
 | `capabilities.sessionResume` | **`true`** | `--resume <session-id>` documented, and every `result` event emits `sessionId`. |
+| `capabilities.resumePromptCarryover` | **`true`** | Command Center resume turns should prepend the browser's recent Q&A transcript because Copilot's session store is opaque to Minions and can resume without enough conversational context. |
 | `capabilities.systemPromptFile` | **`false`** | No `--system-prompt-file` flag exists. Inject system prompt via a `<system>` block prepended to stdin. |
 | `capabilities.effortLevels` | **`true`** | `--effort` accepts `low|medium|high|xhigh` (no `max`). Adapter must map `'max' → 'xhigh'`. |
 | `capabilities.costTracking` | **`false`** | `result.usage` contains `premiumRequests` (count, not USD), no token counts, no cost. |

package/engine/copilot-models.json

@@ -1,5 +1,5 @@
 {
   "runtime": "copilot",
   "models": null,
-  "cachedAt": "2026-05-07T16:58:12.762Z"
+  "cachedAt": "2026-05-07T17:23:11.001Z"
 }

package/engine/runtimes/claude.js

@@ -717,6 +717,8 @@ const capabilities = {
   fallbackModel: true,
   // Engine controls session persistence (writes session.json on completion)
   sessionPersistenceControl: true,
+  // Claude resume reliably restores prior turns; do not duplicate browser transcript.
+  resumePromptCarryover: false,
   // Adapter implements createStreamConsumer(ctx) — required by llm.js accumulator
   streamConsumer: true,
 };

package/engine/runtimes/copilot.js

@@ -874,6 +874,9 @@ const capabilities = {
   fallbackModel: false,
   // Copilot manages session state internally in ~/.copilot/session-state/
   sessionPersistenceControl: false,
+  // CC resumes should include recent visible Q&A in stdin because Minions cannot
+  // inspect or repair Copilot's opaque session-state store when it drops context.
+  resumePromptCarryover: true,
   // Adapter implements createStreamConsumer(ctx) — required by llm.js accumulator
   streamConsumer: true,
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@yemi33/minions",
-  "version": "0.1.1771",
+  "version": "0.1.1773",
   "description": "Multi-agent AI dev team that runs from ~/.minions/ — five autonomous agents share a single engine, dashboard, and knowledge base",
   "bin": {
     "minions": "bin/minions.js"