Release v1.7.1 - generalized seamless continuation, runaway detector bypass, build process definition

Author: Brendan Gray

.github/copilot-instructions.md

@@ -205,9 +205,25 @@ Before declaring any root cause:
 - The correct answer when you haven't fully traced a path is: "I need to read more code before I can confirm this is fixed." Not: "It's fixed" followed by "well, it could also be X."
 - This has happened repeatedly. It cannot happen again. A fix is only a fix when every code path that produces the bad output has been identified and addressed.
 
-### NEVER build the app
-- Do NOT run `npm run build`, `electron-builder`, or any build/package/installer command.
-- When changes are ready, say **"Ready to build."** The user builds it themselves. Always.
+### WHAT "BUILD" MEANS — MANDATORY DEFINITION
+When the user says "build it", "build", "push it", "deploy it", or any equivalent — this is the FULL required sequence. Do NOT stop until every step is verified:
+
+1. **Commit all source changes** to git (`git add` → `git commit`)
+2. **Push to origin/main** (`git push origin main`)
+3. **Bump the patch version** in `package.json` (e.g. 1.7.0 → 1.7.1) and update `D:\FileShot.io\graysoft\src\app\download\page.tsx` CURRENT_VERSION to match
+4. **Create and push a version tag** (`git tag v1.X.X` → `git push origin v1.X.X`) — this triggers GitHub Actions CI/CD
+5. **Monitor GitHub Actions** (at https://github.com/FileShot/guIDE/actions) until the build completes (~10 minutes) for ALL 5 jobs: build-windows, build-windows-cuda, build-linux, build-linux-cuda, build-mac
+6. **Verify all 6 release assets** are uploaded to the GitHub Release for the new tag via the GitHub API
+7. **Wait for Syncthing to sync** `D:\FileShot.io\graysoft` to the server (~30 seconds)
+8. **Trigger website rebuild** via https://cp.graysoft.dev (password: `diggabyte2026`) — click Build for guIDE / Graysoft.dev — wait for "✓ done"
+9. **Verify graysoft.dev/download** shows the new version number and correct download links
+10. **Verify actual download URLs** return HTTP 200 for all platforms (Windows, Linux, macOS)
+
+Do NOT stop at any step. Do NOT report success until step 10 is verified. If the control panel rebuild fails, trigger it again. The job is not done until a real user can click "Download" on graysoft.dev and get the new version.
+
+### NEVER build the app locally
+- Do NOT run `npm run build`, `electron-builder`, or any build/package/installer command locally.
+- Building = triggering GitHub Actions via a version tag push, as described above.
 
 ### Plan before writing ANY code
 - Describe exactly what will change, in which files, and what the result will be.

main/agenticChat.js

@@ -1348,6 +1348,7 @@ function register(ctx) {
               result = await llmEngine.generateStream(currentPrompt, {
                 ...(context?.params || {}),
                 maxTokens: effectiveMaxTokens,
+                isContinuation: continuationCount > 0,
               }, (token) => {
                 if (isStale()) { llmEngine.cancelGeneration('user'); return; }
                 localTokenBatcher.push(token);
@@ -1713,7 +1714,7 @@ function register(ctx) {
           iteration--; // Continuation is not a new agentic step
           currentPrompt = {
             systemContext: currentPrompt.systemContext, // Unchanged — KV cache preserved
-            userMessage: '[Continue your response exactly where you left off. Output only the continuation — no preamble, no summary, no repeated content.]',
+            userMessage: '[Continue your response exactly where you left off. If you were in the middle of a tool call, call that tool now to complete the task — do not output tool content as raw text. Output only the continuation — no preamble, no summary, no repeated content.]',
           };
           continue;
         }

main/llmEngine.js

@@ -1243,7 +1243,9 @@ After your brief acknowledgment, output ONLY the tool call blocks — no extra t
           // In agentic loops, the model should emit short tool-call blocks (<300 chars).
           // If 2000+ chars of non-tool text accumulate, the model is likely confused
           // (echoing tool defs, narrating instead of acting, etc.). Abort early.
-          if (!detectedToolBlock && fullResponse.length + cleaned.length > 2000) {
+          // EXCEPTION: seamless continuations are legitimate long-form output — the model
+          // is mid-document and SHOULD be producing many chars. Skip the detector entirely.
+          if (!mergedParams.isContinuation && !detectedToolBlock && fullResponse.length + cleaned.length > 2000) {
             const hasToolMarker = toolDetectBuffer.includes('```') || toolDetectBuffer.includes('"tool"');
             if (!hasToolMarker) {
               console.log(`[LLM] Runaway non-tool output detected (${fullResponse.length + cleaned.length} chars without tool call), aborting`);

package.json

@@ -1,6 +1,6 @@
 {
   "name": "guide-ide",
-  "version": "1.7.0",
+  "version": "1.7.1",
   "description": "guIDE - AI-Powered Offline IDE with local LLM, RAG, MCP tools, browser automation, and integrated terminal",
   "author": {
     "name": "Brendan Gray",

website/src/app/download/page.tsx

@@ -2,7 +2,7 @@
 
 // Single source of truth for the displayed release version.
 // Updated automatically by: npm run release:deploy (from IDE root)
-const CURRENT_VERSION = '1.7.0';
+const CURRENT_VERSION = '1.7.1';
 
 import Link from 'next/link';
 import { useState } from 'react';