support responses api , support native message-api, fix inconsistent credit consumption in chat

.claude/hooks/subagent-start-marker.js

@@ -0,0 +1,33 @@
+async function readStdin() {
+  let input = "";
+  for await (const chunk of process.stdin) {
+    input += chunk;
+  }
+  return input.trim();
+}
+
+const rawInput = await readStdin();
+let hookInput = {};
+
+if (rawInput) {
+  try {
+    hookInput = JSON.parse(rawInput);
+  } catch {
+    hookInput = {};
+  }
+}
+
+const marker = `__SUBAGENT_MARKER__${JSON.stringify({
+  session_id: hookInput.session_id ?? null,
+  agent_id: hookInput.agent_id ?? null,
+  agent_type: hookInput.agent_type ?? null,
+})}`;
+
+const payload = {
+  hookSpecificOutput: {
+    hookEventName: "SubagentStart",
+    additionalContext: marker,
+  },
+};
+
+process.stdout.write(`${JSON.stringify(payload)}\n`);

.claude/settings.json

@@ -0,0 +1,15 @@
+{
+  "hooks": {
+    "SubagentStart": [
+      {
+        "matcher": "*",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "node --input-type=module -e \"import { homedir } from 'node:os'; import { join } from 'node:path'; import { readFile } from 'node:fs/promises'; const file = join(homedir(), '.claude', 'hooks', 'subagent-start-marker.js'); const source = await readFile(file, 'utf8'); const url = 'data:text/javascript;base64,' + Buffer.from(source).toString('base64'); await import(url);\""
+          }
+        ]
+      }
+    ]
+  }
+}

.opencode/plugins/subagent-marker.js

@@ -0,0 +1,65 @@
+const MARKER_PREFIX = "__SUBAGENT_MARKER__"
+
+const subagentSessions = new Set()
+const markedSessions = new Set()
+
+const getSessionInfo = (event) => {
+  if (!event || typeof event !== "object") return undefined
+  const properties = event.properties
+  if (!properties || typeof properties !== "object") return undefined
+  const info = properties.info
+  if (!info || typeof info !== "object") return undefined
+  return info
+}
+
+export const SubagentMarkerPlugin = async () => {
+  return {
+    event: async ({ event }) => {
+      if (event.type === "session.created") {
+        const info = getSessionInfo(event)
+        if (info?.id && info.parentID) {
+          subagentSessions.add(info.id)
+        }
+        return
+      }
+
+      if (event.type === "session.deleted") {
+        const info = getSessionInfo(event)
+        if (info?.id) {
+          subagentSessions.delete(info.id)
+          markedSessions.delete(info.id)
+        }
+      }
+    },
+    "chat.message": async (input, output) => {
+      const { sessionID } = input
+      if (!subagentSessions.has(sessionID) || markedSessions.has(sessionID)) {
+        return
+      }
+      if (!output.message?.id || !output.message?.sessionID) {
+        return
+      }
+
+      const marker = `${MARKER_PREFIX}${JSON.stringify({
+        session_id: sessionID,
+        agent_id: sessionID,
+        agent_type: input.agent ?? "opencode-subagent",
+      })}`
+
+      output.parts.unshift({
+        id: `${output.message.id}-subagent-marker`,
+        sessionID: output.message.sessionID,
+        messageID: output.message.id,
+        type: "text",
+        text: `<system-reminder>\nSubagentStart hook additional context: ${marker}\n</system-reminder>`,
+        synthetic: true,
+        time: {
+          start: Date.now(),
+          end: Date.now(),
+        },
+      })
+
+      markedSessions.add(sessionID)
+    },
+  }
+}

README.md

@@ -31,7 +31,7 @@ A reverse-engineered proxy for the GitHub Copilot API that exposes it as an Open
 
 ## Features
 
-- **OpenAI & Anthropic Compatibility**: Exposes GitHub Copilot as an OpenAI-compatible (`/v1/chat/completions`, `/v1/models`, `/v1/embeddings`) and Anthropic-compatible (`/v1/messages`) API.
+- **OpenAI & Anthropic Compatibility**: Exposes GitHub Copilot as an OpenAI-compatible (`/v1/responses`, `/v1/chat/completions`, `/v1/models`, `/v1/embeddings`) and Anthropic-compatible (`/v1/messages`) API.
 - **Claude Code Integration**: Easily configure and launch [Claude Code](https://docs.anthropic.com/en/docs/claude-code/overview) to use Copilot as its backend with a simple command-line flag (`--claude-code`).
 - **Usage Dashboard**: A web-based dashboard to monitor your Copilot API usage, view quotas, and see detailed statistics.
 - **Rate Limit Control**: Manage API usage with rate-limiting options (`--rate-limit`) and a waiting mechanism (`--wait`) to prevent errors from rapid requests.
@@ -177,6 +177,52 @@ The following command line options are available for the `start` command:
 | ------ | ------------------------- | ------- | ----- |
 | --json | Output debug info as JSON | false   | none  |
 
+## Configuration (config.json)
+
+- **Location:** `~/.local/share/copilot-api/config.json` (Linux/macOS) or `%USERPROFILE%\.local\share\copilot-api\config.json` (Windows).
+- **Default shape:**
+  ```json
+  {
+    "auth": {
+      "apiKeys": []
+    },
+    "extraPrompts": {
+      "gpt-5-mini": "<built-in exploration prompt>",
+      "gpt-5.1-codex-max": "<built-in exploration prompt>"
+    },
+    "smallModel": "gpt-5-mini",
+    "modelReasoningEfforts": {
+      "gpt-5-mini": "low"
+    },
+    "useFunctionApplyPatch": true,
+    "compactUseSmallModel": true
+  }
+  ```
+- **auth.apiKeys:** API keys used for request authentication. Supports multiple keys for rotation. Requests can authenticate with either `x-api-key: <key>` or `Authorization: Bearer <key>`. If empty or omitted, authentication is disabled.
+- **extraPrompts:** Map of `model -> prompt` appended to the first system prompt when translating Anthropic-style requests to Copilot. Use this to inject guardrails or guidance per model. Missing default entries are auto-added without overwriting your custom prompts.
+- **smallModel:** Fallback model used for tool-less warmup messages (e.g., Claude Code probe requests) to avoid spending premium requests; defaults to `gpt-5-mini`.
+- **modelReasoningEfforts:** Per-model `reasoning.effort` sent to the Copilot Responses API. Allowed values are `none`, `minimal`, `low`, `medium`, `high`, and `xhigh`. If a model isn’t listed, `high` is used by default.
+- **useFunctionApplyPatch:** When `true`, the server will convert any custom tool named `apply_patch` in Responses payloads into an OpenAI-style function tool (`type: "function"`) with a parameter schema so assistants can call it using function-calling semantics to edit files. Set to `false` to leave tools unchanged. Defaults to `true`.
+- **compactUseSmallModel:** When `true`, detected "compact" requests (e.g., from Claude Code or Opencode compact mode) will automatically use the configured `smallModel` to avoid consuming premium model usage for short/background tasks. Defaults to `true`.
+
+Edit this file to customize prompts or swap in your own fast model. Restart the server (or rerun the command) after changes so the cached config is refreshed.
+
+## API Authentication
+
+- **Protected routes:** All routes except `/` require authentication when `auth.apiKeys` is configured and non-empty.
+- **Allowed auth headers:**
+  - `x-api-key: <your_key>`
+  - `Authorization: Bearer <your_key>`
+- **CORS preflight:** `OPTIONS` requests are always allowed.
+- **When no keys are configured:** Server starts normally and allows requests (authentication disabled).
+
+Example request:
+
+```sh
+curl http://localhost:4141/v1/models \
+  -H "x-api-key: your_api_key"
+```
+
 ## API Endpoints
 
 The server exposes several endpoints to interact with the Copilot API. It provides OpenAI-compatible endpoints and now also includes support for Anthropic-compatible endpoints, allowing for greater flexibility with different tools and services.
@@ -185,11 +231,12 @@ The server exposes several endpoints to interact with the Copilot API. It provid
 
 These endpoints mimic the OpenAI API structure.
 
-| Endpoint                    | Method | Description                                               |
-| --------------------------- | ------ | --------------------------------------------------------- |
-| `POST /v1/chat/completions` | `POST` | Creates a model response for the given chat conversation. |
-| `GET /v1/models`            | `GET`  | Lists the currently available models.                     |
-| `POST /v1/embeddings`       | `POST` | Creates an embedding vector representing the input text.  |
+| Endpoint                    | Method | Description                                                      |
+| --------------------------- | ------ | ---------------------------------------------------------------- |
+| `POST /v1/responses`        | `POST` | OpenAI Most advanced interface for generating model responses.          |
+| `POST /v1/chat/completions` | `POST` | Creates a model response for the given chat conversation.        |
+| `GET /v1/models`            | `GET`  | Lists the currently available models.                            |
+| `POST /v1/embeddings`       | `POST` | Creates an embedding vector representing the input text.         |
 
 ### Anthropic Compatible Endpoints
 
@@ -307,12 +354,14 @@ Here is an example `.claude/settings.json` file:
   "env": {
     "ANTHROPIC_BASE_URL": "http://localhost:4141",
     "ANTHROPIC_AUTH_TOKEN": "dummy",
-    "ANTHROPIC_MODEL": "gpt-4.1",
-    "ANTHROPIC_DEFAULT_SONNET_MODEL": "gpt-4.1",
-    "ANTHROPIC_SMALL_FAST_MODEL": "gpt-4.1",
-    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "gpt-4.1",
+    "ANTHROPIC_MODEL": "gpt-5.2",
+    "ANTHROPIC_DEFAULT_SONNET_MODEL": "gpt-5.2",
+    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "gpt-5-mini",
     "DISABLE_NON_ESSENTIAL_MODEL_CALLS": "1",
-    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1"
+    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1",
+    "BASH_MAX_TIMEOUT_MS": "600000",
+    "CLAUDE_CODE_ATTRIBUTION_HEADER": "0",
+    "CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION": "false"
   },
   "permissions": {
     "deny": [
@@ -326,6 +375,45 @@ You can find more options here: [Claude Code settings](https://docs.anthropic.co
 
 You can also read more about IDE integration here: [Add Claude Code to your IDE](https://docs.anthropic.com/en/docs/claude-code/ide-integrations)
 
+### Subagent Marker Integration (Optional)
+
+This project supports `X-Initiator: agent` for subagent-originated requests
+
+#### Claude Code hook producer
+
+Use the included hook script to inject marker context on `SubagentStart`.
+If you place the script under your user Claude directory (`~/.claude/hooks`), use this cross-platform command in `.claude/settings.json`:
+
+- `.claude/hooks/subagent-start-marker.js`
+
+And enable it from `.claude/settings.json`:
+
+```json
+{
+  "hooks": {
+    "SubagentStart": [
+      {
+        "matcher": "*",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "node --input-type=module -e \"import { homedir } from 'node:os'; import { join } from 'node:path'; import { readFile } from 'node:fs/promises'; const file = join(homedir(), '.claude', 'hooks', 'subagent-start-marker.js'); const source = await readFile(file, 'utf8'); const url = 'data:text/javascript;base64,' + Buffer.from(source).toString('base64'); await import(url);\""
+          }
+        ]
+      }
+    ]
+  }
+}
+```
+
+#### Opencode plugin producer
+
+For opencode, use the plugin implementation at:
+
+- `.opencode/plugins/subagent-marker.js`
+
+This plugin tracks sub-sessions and prepends a marker system reminder to subagent chat messages.
+
 ## Running from Source
 
 The project can be run from source in several ways:
@@ -349,3 +437,10 @@ bun run start
   - `--rate-limit <seconds>`: Enforces a minimum time interval between requests. For example, `copilot-api start --rate-limit 30` will ensure there's at least a 30-second gap between requests.
   - `--wait`: Use this with `--rate-limit`. It makes the server wait for the cooldown period to end instead of rejecting the request with an error. This is useful for clients that don't automatically retry on rate limit errors.
 - If you have a GitHub business or enterprise plan account with Copilot, use the `--account-type` flag (e.g., `--account-type business`). See the [official documentation](https://docs.github.com/en/enterprise-cloud@latest/copilot/managing-copilot/managing-github-copilot-in-your-organization/managing-access-to-github-copilot-in-your-organization/managing-github-copilot-access-to-your-organizations-network#configuring-copilot-subscription-based-network-routing-for-your-enterprise-or-organization) for more details.
+
+### CLAUDE.md Recommended Content
+
+Please include the following in `CLAUDE.md` (for Claude usage):
+
+- Prohibited from directly asking questions to users, MUST use AskUserQuestion tool.
+- Once you can confirm that the task is complete, MUST use AskUserQuestion tool to make user confirm. The user may respond with feedback if they are not satisfied with the result, which you can use to make improvements and try again.

eslint.config.js

@@ -1,6 +1,7 @@
 import config from "@echristian/eslint-config"
 
 export default config({
+  ignores: [".claude/**", ".opencode/**"],
   prettier: {
     plugins: ["prettier-plugin-packagejson"],
   },

src/lib/api-config.ts

@@ -7,11 +7,11 @@ export const standardHeaders = () => ({
   accept: "application/json",
 })
 
-const COPILOT_VERSION = "0.26.7"
+const COPILOT_VERSION = "0.37.6"
 const EDITOR_PLUGIN_VERSION = `copilot-chat/${COPILOT_VERSION}`
 const USER_AGENT = `GitHubCopilotChat/${COPILOT_VERSION}`
 
-const API_VERSION = "2025-04-01"
+const API_VERSION = "2025-10-01"
 
 export const copilotBaseUrl = (state: State) =>
   state.accountType === "individual" ?
@@ -25,7 +25,7 @@ export const copilotHeaders = (state: State, vision: boolean = false) => {
     "editor-version": `vscode/${state.vsCodeVersion}`,
     "editor-plugin-version": EDITOR_PLUGIN_VERSION,
     "user-agent": USER_AGENT,
-    "openai-intent": "conversation-panel",
+    "openai-intent": "conversation-agent",
     "x-github-api-version": API_VERSION,
     "x-request-id": randomUUID(),
     "x-vscode-user-agent-library-version": "electron-fetch",