- Home
- Skills
- Technical Documentation
- geminiX voice agent
Works with the AI Agents you already use
geminiX voice agent
Trace, map, and explain the architecture and data flows of the GeminiX VS Code extension codebase.
$6
geminiX voice agent
Example session with this skill installed
Explain how a user message travels from the webview to the Gemini API and back to the UI.
- Read your context and instructions
- Compiled the geminix voice agent
I have traced the message flow from the webview to the host and back. The system uses a bridge pattern where webview/main.ts dispatches events that are handled by src/extension.ts.
GeminiX Message Flow Analysis
| Stage | Responsible Module | Action |
|---|---|---|
| Initiation | webview/main.ts | Captures user input and sends a vscode.postMessage to the host. |
| Routing | src/extension.ts | The onDidReceiveMessage handler identifies the action and routes it to the session manager. |
| Context | src/editorContext.ts | Gathers active editor selection and file metadata to augment the prompt. |
| Protocol | src/liveProtocol.ts | Wraps the message and context into the WebSocket format for the Gemini API. |
| Execution | src/liveSession.ts | Manages the active connection and awaits the streaming response. |
| Streaming | webview/streaming.ts | Receives delta packets from the host and updates the internal state. |
| Rendering | webview/renderCoordinator.ts | Manages the Markdown parser and ensures smooth UI updates during playback. |
Component Relationships
- Data Integrity:
src/chatSchema.tsvalidates the message structure before it is saved tosrc/chatHistory.ts. - State:
src/liveInteraction.tsmaintains the "busy" or "typing" state visible in the UI.
Next steps
- Inspect
src/extension.tsto see the specific switch-case handling for message types. - Review
webview/markdownParser.tsif you need to modify how the response is styled. - Run
npm run test:uito verify the rendering pipeline remains intact.
Connects securely to your tools. The creator never sees your data.
About this skill
Onboarding to a complex VS Code extension codebase is slow and error-prone. This skill automates the discovery process, mapping the specific architecture of GeminiX so you can understand its data flows and module boundaries without getting lost in the source.
What it does
- Architecture mapping: Identifies project structure by analyzing
package.jsonandAGENTS.mdto establish the authoritative tech stack. - Message flow tracing: Tracks communication between the VS Code host (
extension.ts) and the React/Webview frontend (main.ts). - Feature deep-dives: Explains specific subsystems like microphone capture, live transcription, and workspace context retrieval.
- State management analysis: Documents how session state, chat history, and preferences are persisted and validated.
- Verification via tests: Cross-references explanations with existing test suites to confirm edge cases and expected behaviors.
How it works
- Analyze project shape: The agent reads the root configuration to understand scripts, dependencies, and model constraints.
- Identify entry points: It locates the specific feature entry point in the source code based on your query.
- Trace execution paths: It follows data from the UI through host handlers, transformation modules, and persistence layers.
- Deliver evidenced reports: You receive a summary backed by specific file links and symbol references.
Frameworks & tools
This skill is purpose-built for the GeminiX VS Code extension repository. It handles TypeScript, VS Code Extension APIs, WebSocket protocols, and Webview-to-Host communication patterns.
Why this beats prompting it yourself
Generic LLMs often hallucinate file structures or rely on stale documentation. This skill enforces a source-first methodology, tracing actual imports and message types rather than guessing based on variable names.
Use cases
- New contributor onboarding: Quickly understand the extension lifecycle and webview bridge.
- Debugging live sessions: Trace how audio data moves from
microphoneCapture.tstoliveSession.ts. - Context engine auditing: Analyze how
editorContext.tsandworkspaceContext.tsassemble prompts. - UI logic refactoring: Map the rendering pipeline from
streaming.tsto therenderCoordinator.ts.
Known limitations
Focuses exclusively on the GeminiX codebase. Requires local source code access to verify implementation details.
Requires access to the GeminiX source repository. Explanations are limited to what is currently implemented in code and verified by tests.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 15 days ago
Requires
Permissions
Allowed Hosts