ADFA-4589 | Add suite of AI-powered developer plugins for Code on the Go (#41)

* feat: add code-suggestions plugin

Inline ghost-text AI code completions; restored onto main's split layout.

* feat: add speech-to-text plugin

Voice-to-code speech input; restored onto main's split layout.

* feat: add vector-search plugin

Semantic code search via embeddings; restored onto main's split layout.

* fix(ai-core): load the user-selected model instead of newest .gguf in Downloads

Stream the selected content:// URI into a private cache file; drop the Downloads fallback that silently loaded the wrong model (ADFA-4388).

* fix(ai-core): reject embedding models on the chat path instead of crashing

Add GgufModelInspector + UserFeedback; classify the .gguf and throw IncompatibleModelException before native decode SIGABRTs the IDE (ADFA-4388).

* fix(ai-assistant): fetch live Gemini models and drop retired gemini-1.5-flash

listModels now queries the live v1beta ListModels catalog (generateContent-only); default bumped to gemini-2.5-flash and retired models migrated on load.

* fix(ai-assistant): settings dialog UX and backend selection bugs

Readable settings title; real model name shown and persisted across reopen; and the chat backend indicator + routing now follow the selected backend, refreshing when the settings dialog closes.

* fix(code-suggestions-plugin): LLM service lookup

* feat(vector-search): wire semantic project search, also fixed code-suggestion-plugin and vector-search-plugin names

* refactor: Address AI plugin review feedback

- Add version ranges, README/docs, and tiered help coverage for AI plugins
- Cancel plugin coroutine scopes during disposal
- Move Gemini listModels API key from URL query to x-goog-api-key header
- Document Gemini API key storage tradeoff and remove misleading encryption note
- Update AI plugin review status with verification results

* fix(ai-core): shut down Llm-RunLoop executor on plugin dispose

LLamaAndroid's single-thread run-loop executor was never shut down, so the non-daemon Llm-RunLoop thread outlived the plugin and pinned its classloader. Expose LLamaAndroid.shutdown() (closes the ExecutorCoroutineDispatcher) and call it from LocalLlmBackend.close() after the model unload completes. Regenerate the prebuilt llama AAR so the new API is visible to the plugin build.

* refactor: add icons day and night for AI plugins

* fix(ai-assistant): surface LLM-not-configured feedback and follow IDE light/dark theme

- Agent tab now shows an error bubble + Snackbar ("Open Settings...") when no local model or Gemini key is configured, instead of silently doing nothing.
- Theme UI to the IDE day/night setting: route inflation through PluginFragmentHelper.getPluginInflater, make PluginTheme a Material3 DayNight theme, add values-night/ colors, and replace hardcoded hex with @color refs.
- Fix AI Settings toolbar contrast (colorOnPrimary title/icon on the primary bar).

* refactor: extract hardcoded strings to strings.xml and improve user feedback

* fix(ai-plugins): resolve PR review findings across the AI plugin suite

llama run-loop now recreates after per-plugin dispose, embedding-dim guard, synced suggestion cache, capped file walk, fallback-safe Gemini model migration and vector search indexes in the background (bounded wait)

* fix(ai-plugins): honor selected backend and resolve Gemini OkHttp crash

- Resolve backend via "auto" in ai-core
- Migrate Gemini to REST instead of the google-genai SDK to fix OkHttp 3.x conflict
- Implement context-aware suggestion cache key
- Add STT raw-transcript fallback and vector-search roots
- Refactor code: remove unused code and permissions
- Fix version numbers and update plugins documentation

* chore(ai-plugins): align IDE version range, docs, and icons across AI suite

Refresh manifests, README/HTML docs, and day/night icons for ai-core, ai-assistant, code-suggestions, speech-to-text, and vector-search; update PathGuard and search handler.

* fix(ai-core): format local-LLM prompts as ChatML so generation isn't empty

* refactor: update docs for the tab key action and improve prompt for suggestions generation

Author: jatezzz

README.md

@@ -25,6 +25,9 @@ See the official [plugin documentation](https://www.appdevforall.org/codeonthego
 | [`ai-core/`](ai-core/)                             | Shared on-device LLM inference backend (bundled llama.cpp AAR) plus a Gemini API backend, exposed to other plugins as a runtime service. |
 | [`ai-assistant/`](ai-assistant/)                   | In-IDE AI chat assistant with tool calling; talks to `ai-core` for inference over local or Gemini models. |
 | [`flutter-template/`](flutter-template/)           | Adds Flutter starter project templates (Basic, BLoC, Provider, GetX, Riverpod) to the New Project screen. |
+| [`code-suggestions-plugin/`](code-suggestions-plugin/) | Inline ghost-text code completions powered by AI. |
+| [`speech-to-text-plugin/`](speech-to-text-plugin/) | Voice-to-code: converts speech to code with AI generation. |
+| [`vector-search-plugin/`](vector-search-plugin/) | Semantic code search using embeddings and vector similarity. |
 
 ## Building a plugin
 

ai-assistant/README.md

@@ -56,6 +56,22 @@ The build resolves `plugin-api.jar` from the repo-root `../libs/`.
 - `fragments/AiSettingsFragment.kt`, `viewmodel/AiSettingsViewModel.kt` — model/backend config
 - `tool/` — the agent tool-loop (executor, router, per-tool handlers, approval)
 
+## Security
+
+- **Gemini API key at rest.** When you use the Gemini backend, the API key is
+  stored in this plugin's private `SharedPreferences` (`AgentSettings`). That
+  store is **app-sandboxed** — other apps cannot read it — but it is **not
+  encrypted at rest**, so it is recoverable on a rooted or otherwise compromised
+  device. Remove it any time from **AI Settings** (or `clearGeminiApiKey()`).
+  Encryption is deliberately not layered on here: the key is shared at runtime
+  with the sibling `ai-core` plugin (which performs the Gemini calls), and
+  `EncryptedSharedPreferences` would force both independent plugins to agree on a
+  master-key alias and crypto library version — a fragile cross-plugin coupling.
+  A host-provided secure-storage service is the correct long-term fix.
+- **Gemini API key in transit.** The key is sent to Google as an `x-goog-api-key`
+  request header (and via the SDK on the chat path), never in a URL query string.
+- **File tools are confined to the project root** — see the in-IDE help page.
+
 ## License
 
 GPL-3.0 — same as AndroidIDE / CodeOnTheGo.

ai-assistant/ai-assistant.html

@@ -54,9 +54,8 @@ <h2>1. Executive Overview</h2>
     <li><b>AI Core</b> — the inference engine. Provides an
       <code>LlmInferenceService</code> with two backends: a local, on-device
       model (llama.cpp) and Google Gemini. Install this first.</li>
-    <li><b>AI Assistant</b> — the user interface. Contributes the Agent tab, the
-      editor context-menu actions, and the agent tool-loop that drives the
-      model.</li>
+    <li><b>AI Assistant</b> — the user interface. Contributes the Agent tab and
+      the agent tool-loop that drives the model.</li>
   </ul>
   <p>The two plugins are bridged by a shared service registry
   (<code>SharedServices</code>), so the UI plugin discovers the inference
@@ -71,8 +70,6 @@ <h2>2. Core Functionality</h2>
       and generate code from templates.</li>
     <li><b>Dual inference backends</b> — fully offline on-device inference, or
       Gemini in the cloud, selectable in Settings.</li>
-    <li><b>Editor context actions</b> — <b>Explain Code</b> and
-      <b>Generate Code</b> from the editor selection.</li>
     <li><b>Safety controls</b> — filesystem tools are confined to the project
       root, and mutating tools require explicit user approval.</li>
   </ul>

ai-assistant/src/main/AndroidManifest.xml

@@ -25,15 +25,15 @@
 
         <meta-data
             android:name="plugin.author"
-            android:value="AndroidIDE" />
+            android:value="App Dev for All" />
 
         <meta-data
             android:name="plugin.min_ide_version"
             android:value="26.17" />
 
         <meta-data
             android:name="plugin.max_ide_version"
-            android:value="26.28" />
+            android:value="26.30" />
 
         <!-- Number of editor tabs contributed via getEditorTabs() (the "Agent" tab) -->
         <meta-data

ai-assistant/src/main/assets/docs/index.html

@@ -64,10 +64,6 @@ <h2>What the agent can do</h2>
     device.
   </div>
 
-  <h2>Context-menu actions</h2>
-  <p>Select code in the editor and open the context menu for
-  <b>Explain Code</b> and <b>Generate Code</b> shortcuts.</p>
-
   <h2>Troubleshooting</h2>
   <ul>
     <li><b>"No model configured"</b> — select a <code>.gguf</code> file (Local)

ai-assistant/src/main/assets/icon_day.png

@@ -3,7 +3,6 @@ package com.itsaky.androidide.plugins.aiassistant
 import com.itsaky.androidide.plugins.IPlugin
 import com.itsaky.androidide.plugins.PluginContext
 import com.itsaky.androidide.plugins.extensions.UIExtension
-import com.itsaky.androidide.plugins.extensions.ContextMenuContext
 import com.itsaky.androidide.plugins.extensions.DocumentationExtension
 import com.itsaky.androidide.plugins.extensions.MenuItem
 import com.itsaky.androidide.plugins.extensions.PluginTooltipButton
@@ -20,8 +19,20 @@ class AiAssistantPlugin : IPlugin, UIExtension, DocumentationExtension {
     private var llmService: LlmInferenceService? = null
 
     companion object {
+        /** Must match `plugin.id` in AndroidManifest.xml — keys the host's plugin Context lookup
+         *  used by [com.itsaky.androidide.plugins.base.PluginFragmentHelper.getPluginInflater]. */
+        const val PLUGIN_ID = "com.itsaky.androidide.plugins.aiassistant"
+
+        /** Tooltip category for this plugin (strict `plugin_<pluginId>` convention); shared by the tab and the AI Settings screen. */
+        const val TOOLTIP_CATEGORY = "plugin_$PLUGIN_ID"
+
         const val TOOLTIP_TAG_TAB = "agent_chat_tab"
 
+        // Tags for the interactive controls on the AI Settings dialog (see AiSettingsFragment).
+        const val TOOLTIP_TAG_SETTINGS_BACKEND = "ai_settings_backend"
+        const val TOOLTIP_TAG_SETTINGS_LOCAL_MODEL = "ai_settings_local_model"
+        const val TOOLTIP_TAG_SETTINGS_GEMINI_KEY = "ai_settings_gemini_key"
+
         @Volatile
         private var pluginContext: PluginContext? = null
 
@@ -87,30 +98,6 @@ class AiAssistantPlugin : IPlugin, UIExtension, DocumentationExtension {
         )
     }
 
-    override fun getContextMenuItems(menuContext: ContextMenuContext): List<MenuItem> {
-        val selectedText = menuContext.selectedText
-        if (selectedText.isNullOrBlank()) {
-            return emptyList()
-        }
-
-        return listOf(
-            MenuItem(
-                id = "ai_explain_code",
-                title = "Explain Code",
-                isEnabled = true,
-                isVisible = true,
-                action = { context.logger.info("Explain Code clicked") }
-            ),
-            MenuItem(
-                id = "ai_generate_code",
-                title = "Generate Code",
-                isEnabled = true,
-                isVisible = true,
-                action = { context.logger.info("Generate Code clicked") }
-            )
-        )
-    }
-
     override fun getMainMenuItems(): List<MenuItem> = emptyList()
 
     // --- DocumentationExtension: three-tier tooltip help for the Agent tab ---
@@ -120,7 +107,7 @@ class AiAssistantPlugin : IPlugin, UIExtension, DocumentationExtension {
     //   Tier 3 = `buttons[].uri`  (offline HTML page served from
     //                              src/main/assets/docs/ at localhost)
 
-    override fun getTooltipCategory(): String = "plugin_ai_assistant"
+    override fun getTooltipCategory(): String = TOOLTIP_CATEGORY
 
     override fun getTooltipEntries(): List<PluginTooltipEntry> = listOf(
         PluginTooltipEntry(
@@ -147,6 +134,50 @@ class AiAssistantPlugin : IPlugin, UIExtension, DocumentationExtension {
                     order = 0
                 )
             )
+        ),
+        PluginTooltipEntry(
+            tag = TOOLTIP_TAG_SETTINGS_BACKEND,
+            summary = "Choose which model powers the Agent: on-device Local (llama.cpp) or cloud Gemini.",
+            detail = """
+                <p>Selects the active inference backend:</p>
+                <ul>
+                  <li><b>Local</b> — runs a <code>.gguf</code> model entirely on
+                      the device; nothing leaves the phone.</li>
+                  <li><b>Gemini</b> — calls Google's cloud API over HTTPS; needs
+                      an API key.</li>
+                </ul>
+                <p>The choice below changes which settings appear.</p>
+            """.trimIndent(),
+            buttons = listOf(
+                PluginTooltipButton(description = "AI Assistant guide", uri = "index.html", order = 0)
+            )
+        ),
+        PluginTooltipEntry(
+            tag = TOOLTIP_TAG_SETTINGS_LOCAL_MODEL,
+            summary = "Pick a local .gguf chat model to run on-device.",
+            detail = """
+                <p>Browse for a <code>.gguf</code> model file to load with
+                llama.cpp. Use a <b>chat/instruct</b> model — embedding-only
+                models can't generate replies. Larger models are slower and use
+                more memory; the file is copied into the app's private storage on
+                first use.</p>
+            """.trimIndent(),
+            buttons = listOf(
+                PluginTooltipButton(description = "AI Assistant guide", uri = "index.html", order = 0)
+            )
+        ),
+        PluginTooltipEntry(
+            tag = TOOLTIP_TAG_SETTINGS_GEMINI_KEY,
+            summary = "Enter your Google Gemini API key. It is stored only on this device.",
+            detail = """
+                <p>Paste a Gemini API key to enable the cloud backend. The key is
+                kept in this plugin's private preferences on-device and is sent
+                only to Google's API over HTTPS. Requests (your prompts and
+                project context) leave the device when Gemini is selected.</p>
+            """.trimIndent(),
+            buttons = listOf(
+                PluginTooltipButton(description = "AI Assistant guide", uri = "index.html", order = 0)
+            )
         )
     )
 
@@ -229,8 +260,6 @@ class AiAssistantPlugin : IPlugin, UIExtension, DocumentationExtension {
                 }
             }
 
-            // Note: Encrypted Gemini API key migration handled by EncryptedPrefs
-
             if (migratedCount > 0) {
                 context.logger.info("Migrated $migratedCount settings from app to plugin")
             } else {

ai-assistant/src/main/assets/icon_night.png

@@ -27,7 +27,6 @@ import java.util.Date
 import java.util.Locale
 
 class ChatAdapter(
-    private val pluginContext: Context,
     private val markwon: Markwon,
     private val onMessageAction: (action: String, message: ChatMessage) -> Unit
 ) : ListAdapter<ChatMessage, RecyclerView.ViewHolder>(DiffCallback) {
@@ -84,8 +83,8 @@ class ChatAdapter(
 
     override fun onCreateViewHolder(parent: ViewGroup, viewType: Int): RecyclerView.ViewHolder {
         android.util.Log.d("ChatAdapter", "onCreateViewHolder called, viewType=$viewType")
-        // Use plugin context for inflating layouts to access plugin resources
-        val inflater = LayoutInflater.from(pluginContext)
+        // Inflate from the RecyclerView's Context so item views follow the IDE day/night theme.
+        val inflater = LayoutInflater.from(parent.context)
         return when (viewType) {
             VIEW_TYPE_SYSTEM -> {
                 val view = inflater.inflate(R.layout.list_item_chat_system_message, parent, false)