@leo000001/claude-code-mcp 1.6.0 → 2.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,11 +1,35 @@
1
1
  # Changelog
2
2
 
3
- ## 1.5.0 (2026-02-12)
3
+ ## Unreleased
4
+
5
+ ### Breaking Changes
6
+ - `claude_code` and `claude_code_reply` now start asynchronously and return `{ sessionId, status: "running", pollInterval }`. Use `claude_code_check` to poll events and fetch the final `result`.
7
+ - Removed tool: `claude_code_configure`
8
+ - New tool: `claude_code_check` (poll + respond_permission)
9
+ - **Parameter nesting refactor**: low-frequency parameters have been folded into nested objects to reduce top-level clutter. This is a breaking change for callers that pass these parameters at the top level:
10
+ - `claude_code`: 22 low-frequency params moved into `advanced` object (e.g. `effort` → `advanced.effort`, `tools` → `advanced.tools`, `agents` → `advanced.agents`, `env` → `advanced.env`)
11
+ - `claude_code_reply`: 28 disk-resume params moved into `diskResumeConfig` object (e.g. `resumeToken` → `diskResumeConfig.resumeToken`, `cwd` → `diskResumeConfig.cwd`)
12
+ - `claude_code_check`: 9 poll control params moved into `pollOptions` object (e.g. `includeTools` → `pollOptions.includeTools`); 2 permission response params moved into `permissionOptions` object (e.g. `updatedInput` → `permissionOptions.updatedInput`)
13
+
14
+ ### Features
15
+ - New module: `src/tools/query-consumer.ts` — shared background query consumer (`consumeQuery`) for start, resume, and disk-resume code paths
16
+ - New module: `src/tools/tool-discovery.ts` — runtime tool discovery with `TOOL_CATALOG`, `ToolDiscoveryCache`, and dynamic `claude_code` description generation
17
+ - New module: `src/utils/build-options.ts` — centralized SDK `Partial<Options>` construction from flat input objects
18
+ - New module: `src/utils/race-with-abort.ts` — race a promise against an AbortSignal with cleanup
19
+ - New module: `src/utils/resume-token.ts` — HMAC-SHA256 resume token generation/validation for secure disk resume
20
+
21
+ ### Improvements
22
+ - `claude_code_check`: default `responseMode="minimal"` to reduce payload size; supports `maxEvents` pagination with `truncated`/`truncatedFields`
23
+ - `claude_code_check`: minimal mode now slims assistant message events (strips `usage`, `model`, `id`, `cache_control` from content blocks)
24
+ - `claude_code_check`: minimal mode filters out noisy progress events (`tool_progress`, `auth_status`); use `includeProgressEvents: true` to restore
25
+ - `claude_code_check`: minimal mode omits `lastEventId`/`lastToolUseId` from top-level response and `durationApiMs`/`sessionTotalTurns`/`sessionTotalCostUsd` from AgentResult
26
+ - `claude_code_check`: includes lightweight session diagnostics (`cancelledAt`/`cancelledReason`/`cancelledSource`, `lastEventId`, `lastToolUseId`)
27
+ - Disk resume security: disk resume fallback requires `CLAUDE_CODE_MCP_RESUME_SECRET` + `resumeToken`
28
+
29
+ ## 1.6.0 (2026-02-12)
4
30
 
5
31
  ### Bug Fixes
6
32
  - Windows: fixed Git Bash auto-detection path derivation and improved candidate search
7
- - Server: internal exceptions in `claude_code`/`claude_code_reply` now return JSON-shaped error results (consistent parsing)
8
- - Server: tightened input validation for `prompt`, `sessionId`, and `cwd` (trim + non-empty)
9
33
 
10
34
  ### Security
11
35
  - `claude_code_session includeSensitive=true` no longer leaks extra fields (e.g. `env`, `debugFile`, `mcpServers`, `sandbox`) beyond the documented sensitive set
package/README.md CHANGED
@@ -10,15 +10,15 @@ Inspired by the [Codex MCP](https://developers.openai.com/codex/guides/agents-sd
10
10
 
11
11
  ## Features
12
12
 
13
- - **4 tools** covering the full agent lifecycle: start, continue, manage, configure
13
+ - **4 tools** covering the full agent lifecycle: start, continue, check/poll, manage
14
14
  - **Session management** with resume and fork support
15
15
  - **Local settings loaded by default** — automatically reads `~/.claude/settings.json`, `.claude/settings.json`, `.claude/settings.local.json`, and `CLAUDE.md` so the agent behaves like your local Claude Code CLI
16
- - **Fine-grained permissions** — tool allow/deny lists, permission modes
16
+ - **Async permissions** — allow/deny lists + explicit approvals via `claude_code_check`
17
17
  - **Custom subagents** — define specialized agents per session
18
18
  - **Cost tracking** — per-session turn and cost accounting
19
19
  - **Session cancellation** via AbortController
20
20
  - **Auto-cleanup** — 30-minute idle timeout for expired sessions
21
- - **Security** — `bypassPermissions` disabled by default
21
+ - **Security** — callers control tool permissions via allow/deny lists + explicit permission decisions
22
22
 
23
23
  ## Prerequisites
24
24
 
@@ -28,7 +28,7 @@ This MCP server uses the [`@anthropic-ai/claude-agent-sdk`](https://www.npmjs.co
28
28
 
29
29
  - The SDK's bundled CLI version is determined by the SDK package version (e.g. SDK 0.2.38 = Claude Code 2.1.38)
30
30
  - **Configuration is shared** — the bundled CLI reads API keys and settings from `~/.claude/`, same as the system-installed `claude`
31
- - **All local settings are loaded by default** — unlike the raw SDK (which defaults to isolation mode), this MCP server loads `user`, `project`, and `local` settings automatically, including `CLAUDE.md` project context. Pass `settingSources: []` to opt out
31
+ - **All local settings are loaded by default** — unlike the raw SDK (which defaults to isolation mode), this MCP server loads `user`, `project`, and `local` settings automatically, including `CLAUDE.md` project context. Pass `advanced.settingSources: []` to opt out
32
32
  - You must have Claude Code configured (API key set up) before using this MCP server: see [Claude Code documentation](https://docs.anthropic.com/en/docs/claude-code/overview)
33
33
 
34
34
  > **Note:** The bundled CLI version may differ from your system-installed `claude`. To check: `claude --version` (system) vs `npm ls @anthropic-ai/claude-agent-sdk` (SDK).
@@ -88,97 +88,120 @@ npm start
88
88
 
89
89
  Start a new Claude Code session. The agent autonomously performs coding tasks: reading/writing files, running shell commands, searching code, managing git, and interacting with APIs.
90
90
 
91
- | Parameter | Type | Required | Description |
92
- | ---------------------------- | ------------------ | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------- |
93
- | `prompt` | string | Yes | Task or question for Claude Code |
94
- | `cwd` | string | No | Working directory (defaults to server cwd) |
95
- | `allowedTools` | string[] | No | List of tool names the agent can use without permission prompts. In `"dontAsk"` mode, only tools in this list are available. Example: `["Bash", "Read", "Write", "Edit"]` |
96
- | `disallowedTools` | string[] | No | List of tool names the agent is forbidden from using. Takes precedence over `allowedTools` |
97
- | `tools` | string[] \| object | No | Define the base tool set. Array of tool name strings, or `{ type: "preset", preset: "claude_code" }` for the default toolset. `allowedTools`/`disallowedTools` further filter on top of this |
98
- | `persistSession` | boolean | No | Persist session history to disk (`~/.claude/projects/`). Default: `true`. Set `false` to disable. |
99
- | `permissionMode` | string | No | Controls how the agent handles tool permissions. Defaults to `"dontAsk"`. Options: `"default"`, `"acceptEdits"`, `"bypassPermissions"`, `"plan"`, `"delegate"`, `"dontAsk"` |
100
- | `maxTurns` | number | No | Maximum number of agent reasoning steps. Each step may involve one or more tool calls |
101
- | `model` | string | No | Model to use (e.g. `"claude-sonnet-4-5-20250929"`) |
102
- | `systemPrompt` | string \| object | No | Override the agent's system prompt. Pass a string for full replacement, or `{ type: "preset", preset: "claude_code", append?: "..." }` to extend the default prompt |
103
- | `agents` | object | No | Define custom sub-agents the main agent can delegate tasks to. Each key is the agent name; value specifies prompt, tools, model, etc. |
104
- | `maxBudgetUsd` | number | No | Maximum budget in USD |
105
- | `timeout` | number | No | Timeout in milliseconds for this session |
106
- | `effort` | string | No | Effort level: `"low"`, `"medium"`, `"high"`, `"max"` |
107
- | `betas` | string[] | No | Beta features (e.g. `["context-1m-2025-08-07"]`) |
108
- | `additionalDirectories` | string[] | No | Additional directories the agent can access beyond cwd |
109
- | `outputFormat` | object | No | Structured output: `{ type: "json_schema", schema: {...} }`. Omit for plain text |
110
- | `thinking` | object | No | Thinking mode: `{ type: "adaptive" }`, `{ type: "enabled", budgetTokens: N }`, or `{ type: "disabled" }` |
111
- | `pathToClaudeCodeExecutable` | string | No | Path to a custom Claude Code executable |
112
- | `agent` | string | No | Name of a custom agent (defined in `agents`) to use as the primary agent, applying its system prompt, tool restrictions, and model |
113
- | `mcpServers` | object | No | MCP server configurations (key: server name, value: server config) |
114
- | `sandbox` | object | No | Sandbox configuration for isolating shell command execution (e.g., Docker container settings) |
115
- | `fallbackModel` | string | No | Fallback model if the primary model fails or is unavailable |
116
- | `enableFileCheckpointing` | boolean | No | Enable file checkpointing to track file changes during the session |
117
- | `includePartialMessages` | boolean | No | When true, includes intermediate streaming messages in the response. Useful for real-time progress monitoring. Default: false |
118
- | `strictMcpConfig` | boolean | No | Enforce strict validation of MCP server configurations |
119
- | `settingSources` | string[] | No | Which filesystem settings to load. Defaults to `["user", "project", "local"]` (loads all settings and CLAUDE.md). Pass `[]` for SDK isolation mode |
120
- | `debug` | boolean | No | Enable debug mode for verbose logging |
121
- | `debugFile` | string | No | Write debug logs to a specific file path (implicitly enables debug mode) |
122
- | `env` | object | No | Environment variables passed to the Claude Code process |
123
-
124
- **Returns:** `{ sessionId, result, isError, durationMs, durationApiMs?, numTurns, totalCostUsd, sessionTotalTurns?, sessionTotalCostUsd?, structuredOutput?, stopReason?, errorSubtype?, usage?, modelUsage?, permissionDenials? }`
91
+ | Parameter | Type | Required | Description |
92
+ | ---------------------------- | ------------------ | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
93
+ | `prompt` | string | Yes | Task or question for Claude Code |
94
+ | `cwd` | string | No | Working directory (defaults to server cwd) |
95
+ | `allowedTools` | string[] | No | Auto-approved tool names. Default: `[]` (none). Tools not in `allowedTools`/`disallowedTools` may surface permission requests via `claude_code_check`. Example: `["Bash", "Read", "Write", "Edit"]` |
96
+ | `disallowedTools` | string[] | No | Forbidden tool names. Default: `[]` (none). SDK behavior: disallowed tools are removed from the model's context. Takes precedence over `allowedTools` and will be denied even if later approved interactively |
97
+ | `maxTurns` | number | No | Maximum number of agent reasoning steps. Each step may involve one or more tool calls. Default: SDK/Claude Code default |
98
+ | `model` | string | No | Model to use (e.g. `"claude-sonnet-4-5-20250929"`). Default: SDK/Claude Code default |
99
+ | `systemPrompt` | string \| object | No | Override the agent's system prompt. Default: SDK/Claude Code default. Pass a string for full replacement, or `{ type: "preset", preset: "claude_code", append?: "..." }` to extend the default prompt |
100
+ | `permissionRequestTimeoutMs` | number | No | Timeout in milliseconds waiting for permission decisions. Default: `60000` |
101
+ | `advanced` | object | No | Advanced/low-frequency parameters (see below) |
102
+
103
+ <details>
104
+ <summary><code>advanced</code> object parameters (22 low-frequency parameters)</summary>
105
+
106
+ | Parameter | Type | Description |
107
+ | ---------------------------------- | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
108
+ | `advanced.tools` | string[] \| object | Define the base tool set. Default: SDK/Claude Code default toolset. Array of tool name strings, or `{ type: "preset", preset: "claude_code" }` for the default toolset. `allowedTools`/`disallowedTools` further filter on top of this |
109
+ | `advanced.persistSession` | boolean | Persist session history to disk (`~/.claude/projects/`). Default: `true`. Set `false` to disable. |
110
+ | `advanced.sessionInitTimeoutMs` | number | Timeout in milliseconds waiting for `system/init`. Default: `10000` |
111
+ | `advanced.agents` | object | Define custom sub-agents the main agent can delegate tasks to. Default: none. SDK default: if a sub-agent omits `tools`, it inherits all tools from the parent. |
112
+ | `advanced.agent` | string | Name of a custom agent (defined in `agents`) to use as the primary agent. Default: omitted |
113
+ | `advanced.maxBudgetUsd` | number | Maximum budget in USD. Default: SDK/Claude Code default |
114
+ | `advanced.effort` | string | Effort level: `"low"`, `"medium"`, `"high"`, `"max"`. Default: SDK/Claude Code default |
115
+ | `advanced.betas` | string[] | Beta features (e.g. `["context-1m-2025-08-07"]`). Default: none |
116
+ | `advanced.additionalDirectories` | string[] | Additional directories the agent can access beyond cwd. Default: none |
117
+ | `advanced.outputFormat` | object | Structured output: `{ type: "json_schema", schema: {...} }`. Default: omitted (plain text) |
118
+ | `advanced.thinking` | object | Thinking mode: `{ type: "adaptive" }`, `{ type: "enabled", budgetTokens: N }`, or `{ type: "disabled" }`. Default: SDK/Claude Code default |
119
+ | `advanced.pathToClaudeCodeExecutable` | string | Path to the Claude Code executable. Default: SDK-bundled Claude Code (cli.js) |
120
+ | `advanced.mcpServers` | object | MCP server configurations (key: server name, value: server config). Default: none |
121
+ | `advanced.sandbox` | object | Sandbox configuration for isolating shell command execution (e.g., Docker container settings). Default: SDK/Claude Code default |
122
+ | `advanced.fallbackModel` | string | Fallback model if the primary model fails or is unavailable. Default: none |
123
+ | `advanced.enableFileCheckpointing` | boolean | Enable file checkpointing to track file changes during the session. Default: `false` |
124
+ | `advanced.includePartialMessages` | boolean | When true, includes intermediate streaming messages in the response. Useful for real-time progress monitoring. Default: false |
125
+ | `advanced.strictMcpConfig` | boolean | Enforce strict validation of MCP server configurations. Default: `false` |
126
+ | `advanced.settingSources` | string[] | Which filesystem settings to load. Defaults to `["user", "project", "local"]` (loads all settings and CLAUDE.md). Pass `[]` for SDK isolation mode |
127
+ | `advanced.debug` | boolean | Enable debug mode for verbose logging. Default: `false` |
128
+ | `advanced.debugFile` | string | Write debug logs to a specific file path (implicitly enables debug mode). Default: omitted |
129
+ | `advanced.env` | object | Environment variables to merge with process.env and pass to the Claude Code process (user values take precedence). Default: inherit process.env |
130
+
131
+ </details>
132
+
133
+ **Returns:** `{ sessionId, status: "running", pollInterval, resumeToken? }`
134
+
135
+ Notes:
136
+ - `resumeToken` is omitted by default, and is only returned when `CLAUDE_CODE_MCP_RESUME_SECRET` is set on the server.
137
+ - On error: `{ sessionId: "", status: "error", error }`
138
+
139
+ Use `claude_code_check` to poll events and obtain the final `result`.
125
140
 
126
141
  > Notes:
127
142
  > - **Subagents require the `Task` tool** to be available to the primary agent. If you use `allowedTools`, include `"Task"` or the agent will be unable to invoke subagents.
128
- > - If you configure `mcpServers` and want the agent to call tools from those servers, you must also allow them via `allowedTools` (e.g. `"mcp__my_server__*"` or specific tool names), especially in `permissionMode="dontAsk"`.
129
- > - `includePartialMessages` affects the underlying SDK event stream, but **this MCP server returns a single final JSON result** (it does not stream intermediate events over MCP responses).
143
+ > - If you configure `advanced.mcpServers` and want the agent to auto-use tools from those servers without approvals, include the exact tool names in `allowedTools` (e.g. `["mcp__my_server__tools/list"]`). Otherwise you will see permission requests via `claude_code_check`.
144
+ > - `advanced.includePartialMessages` affects the underlying SDK event stream; intermediate messages are captured as events and returned via `claude_code_check` (the `claude_code` call itself does not stream).
130
145
 
131
146
  ### `claude_code_reply` — Continue a session
132
147
 
133
148
  Continue an existing session by sending a follow-up message. The agent retains full context from previous turns including files read, code analysis, and conversation history.
134
149
 
135
- | Parameter | Type | Required | Description |
136
- | ------------- | ------- | -------- | --------------------------------------------- |
137
- | `sessionId` | string | Yes | Session ID from a previous `claude_code` call |
138
- | `prompt` | string | Yes | Follow-up prompt |
139
- | `forkSession` | boolean | No | Create a branched copy of this session. The original remains unchanged; the new session diverges from this point |
140
- | `timeout` | number | No | Timeout in milliseconds for this reply |
150
+ | Parameter | Type | Required | Description |
151
+ | ---------------------------- | ------- | -------- | ---------------------------------------------------------------------------------------------------------------- |
152
+ | `sessionId` | string | Yes | Session ID from a previous `claude_code` call |
153
+ | `prompt` | string | Yes | Follow-up prompt |
154
+ | `forkSession` | boolean | No | Create a branched copy of this session. Default: `false` |
155
+ | `permissionRequestTimeoutMs` | number | No | Timeout in milliseconds waiting for permission decisions. Default: `60000` |
156
+ | `sessionInitTimeoutMs` | number | No | Timeout in milliseconds waiting for fork `system/init`. Default: `10000` |
157
+ | `diskResumeConfig` | object | No | Disk resume parameters (see below). Used when `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1` and in-memory session is missing |
141
158
 
142
159
  <details>
143
- <summary>Disk resume parameters (used when <code>CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1</code> and in-memory session is missing)</summary>
144
-
145
- | Parameter | Type | Description |
146
- | ---------------------------- | ------------------ | --------------------------------------------------- |
147
- | `cwd` | string | Working directory |
148
- | `allowedTools` | string[] | Auto-approved tool names (see `claude_code` tool) |
149
- | `disallowedTools` | string[] | Forbidden tool names (see `claude_code` tool) |
150
- | `tools` | string[] \| object | Base tool set (see `claude_code` tool) |
151
- | `persistSession` | boolean | Persist session history to disk |
152
- | `permissionMode` | string | Permission mode |
153
- | `maxTurns` | number | Maximum number of agent reasoning steps |
154
- | `model` | string | Model to use |
155
- | `systemPrompt` | string \| object | Override the agent's system prompt |
156
- | `agents` | object | Custom sub-agent definitions (see `claude_code`) |
157
- | `maxBudgetUsd` | number | Maximum budget in USD |
158
- | `effort` | string | Effort level |
159
- | `betas` | string[] | Beta features |
160
- | `additionalDirectories` | string[] | Additional directories |
161
- | `outputFormat` | object | Structured output format |
162
- | `thinking` | object | Thinking mode |
163
- | `resumeSessionAt` | string | Resume up to a specific message UUID |
164
- | `pathToClaudeCodeExecutable` | string | Path to Claude Code executable |
165
- | `agent` | string | Primary agent name (see `claude_code` tool) |
166
- | `mcpServers` | object | MCP server configurations |
167
- | `sandbox` | object | Sandbox config for command isolation |
168
- | `fallbackModel` | string | Fallback model |
169
- | `enableFileCheckpointing` | boolean | Enable file checkpointing |
170
- | `includePartialMessages` | boolean | Include intermediate streaming messages |
171
- | `strictMcpConfig` | boolean | Strict MCP config validation |
172
- | `settingSources` | string[] | Which filesystem settings to load (defaults to all) |
173
- | `debug` | boolean | Debug mode |
174
- | `debugFile` | string | Debug log file path |
175
- | `env` | object | Environment variables |
160
+ <summary><code>diskResumeConfig</code> object parameters (28 disk-resume-only parameters)</summary>
161
+
162
+ | Parameter | Type | Description |
163
+ | -------------------------------------------- | ------------------ | --------------------------------------------------- |
164
+ | `diskResumeConfig.resumeToken` | string | Resume token returned by `claude_code` / `claude_code_reply`. Required for disk resume fallback |
165
+ | `diskResumeConfig.cwd` | string | Working directory. Required for disk resume. |
166
+ | `diskResumeConfig.allowedTools` | string[] | Auto-approved tool names (see `claude_code`). Default: `[]` |
167
+ | `diskResumeConfig.disallowedTools` | string[] | Forbidden tool names (see `claude_code`). Default: `[]` |
168
+ | `diskResumeConfig.tools` | string[] \| object | Base tool set (see `claude_code`). Default: SDK/Claude Code default |
169
+ | `diskResumeConfig.persistSession` | boolean | Persist session history to disk. Default: `true` |
170
+ | `diskResumeConfig.maxTurns` | number | Maximum number of agent reasoning steps. Default: SDK/Claude Code default |
171
+ | `diskResumeConfig.model` | string | Model to use. Default: SDK/Claude Code default |
172
+ | `diskResumeConfig.systemPrompt` | string \| object | Override the agent's system prompt. Default: SDK/Claude Code default |
173
+ | `diskResumeConfig.agents` | object | Custom sub-agent definitions (see `claude_code`). Default: none |
174
+ | `diskResumeConfig.agent` | string | Primary agent name (see `claude_code` tool). Default: omitted |
175
+ | `diskResumeConfig.maxBudgetUsd` | number | Maximum budget in USD. Default: SDK/Claude Code default |
176
+ | `diskResumeConfig.effort` | string | Effort level. Default: SDK/Claude Code default |
177
+ | `diskResumeConfig.betas` | string[] | Beta features. Default: none |
178
+ | `diskResumeConfig.additionalDirectories` | string[] | Additional directories. Default: none |
179
+ | `diskResumeConfig.outputFormat` | object | Structured output format. Default: omitted (plain text) |
180
+ | `diskResumeConfig.thinking` | object | Thinking mode. Default: SDK/Claude Code default |
181
+ | `diskResumeConfig.resumeSessionAt` | string | Resume only up to and including a specific message UUID. Default: omitted |
182
+ | `diskResumeConfig.pathToClaudeCodeExecutable` | string | Path to Claude Code executable. Default: SDK-bundled Claude Code (cli.js) |
183
+ | `diskResumeConfig.mcpServers` | object | MCP server configurations. Default: none |
184
+ | `diskResumeConfig.sandbox` | object | Sandbox config for command isolation. Default: SDK/Claude Code default |
185
+ | `diskResumeConfig.fallbackModel` | string | Fallback model. Default: none |
186
+ | `diskResumeConfig.enableFileCheckpointing` | boolean | Enable file checkpointing. Default: `false` |
187
+ | `diskResumeConfig.includePartialMessages` | boolean | Include intermediate streaming messages. Default: `false` |
188
+ | `diskResumeConfig.strictMcpConfig` | boolean | Strict MCP config validation. Default: `false` |
189
+ | `diskResumeConfig.settingSources` | string[] | Which filesystem settings to load. Default: `["user", "project", "local"]` |
190
+ | `diskResumeConfig.debug` | boolean | Debug mode. Default: `false` |
191
+ | `diskResumeConfig.debugFile` | string | Debug log file path. Default: omitted |
192
+ | `diskResumeConfig.env` | object | Environment variables. Default: inherit process.env (user values override) |
176
193
 
177
194
  </details>
178
195
 
179
- **Returns:** `{ sessionId, result, isError, durationMs, durationApiMs?, numTurns, totalCostUsd, sessionTotalTurns?, sessionTotalCostUsd?, structuredOutput?, stopReason?, errorSubtype?, usage?, modelUsage?, permissionDenials? }`
196
+ **Returns:** `{ sessionId, status: "running", pollInterval, resumeToken? }`
180
197
 
181
- **Disk resume (optional):** By default, `claude_code_reply` requires the session to exist in the MCP server's in-memory Session Manager. If you set `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1`, it will attempt to resume using the Claude Code CLI's on-disk transcript even when the in-memory session is missing (e.g. after a restart / TTL cleanup). In that mode, you may also pass the session options listed in the collapsible table above, which are otherwise ignored when the in-memory session exists.
198
+ Notes:
199
+ - `resumeToken` is omitted by default, and is only returned when `CLAUDE_CODE_MCP_RESUME_SECRET` is set on the server.
200
+ - On error: `{ sessionId, status: "error", error }`
201
+
202
+ Use `claude_code_check` to poll events and obtain the final `result`.
203
+
204
+ **Disk resume (optional):** By default, `claude_code_reply` requires the session to exist in the MCP server's in-memory Session Manager. If you set `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1`, it can attempt to resume using the Claude Code CLI's on-disk transcript even when the in-memory session is missing (e.g. after a restart / TTL cleanup). For safety, disk resume fallback requires `CLAUDE_CODE_MCP_RESUME_SECRET` to be set on the server and requires callers to pass `diskResumeConfig.resumeToken` (returned by `claude_code` / `claude_code_reply` when `CLAUDE_CODE_MCP_RESUME_SECRET` is set).
182
205
 
183
206
  ### `claude_code_session` — Manage sessions
184
207
 
@@ -188,48 +211,112 @@ List, inspect, or cancel sessions.
188
211
  | ------------------ | ------- | -------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
189
212
  | `action` | string | Yes | `"list"`, `"get"`, or `"cancel"` |
190
213
  | `sessionId` | string | For get/cancel | Target session ID |
191
- | `includeSensitive` | boolean | No | Include `cwd`/`systemPrompt`/`agents`/`additionalDirectories` (default: false; requires `CLAUDE_CODE_MCP_ALLOW_SENSITIVE_SESSION_DETAILS=1`) |
214
+ | `includeSensitive` | boolean | No | Include `cwd`/`systemPrompt`/`agents`/`additionalDirectories` (default: false) |
192
215
 
193
216
  **Returns:** `{ sessions, message?, isError? }`
194
217
 
195
- ### `claude_code_configure` — Runtime configuration
218
+ ### `claude_code_check` — Poll events and respond to permission requests
219
+
220
+ Poll session events/results and approve/deny pending permission requests.
221
+
222
+ | Parameter | Type | Required | Description |
223
+ | ------------------------- | ------- | ---------------------- | -------------------------------------------------------------------------------------------------------------------------- |
224
+ | `action` | string | Yes | `"poll"` or `"respond_permission"` |
225
+ | `sessionId` | string | Yes | Target session ID |
226
+ | `cursor` | number | No | Event cursor for incremental polling (`poll` only). Default: omitted (starts from the beginning of the buffer) |
227
+ | `responseMode` | string | No | `"minimal"` (default) or `"full"` — controls payload size and redaction behavior |
228
+ | `maxEvents` | number | No | Max events per poll (pagination via `nextCursor`). Default: `200` in `"minimal"`; unlimited in `"full"` |
229
+ | `requestId` | string | For respond_permission | Permission request ID |
230
+ | `decision` | string | For respond_permission | `"allow"` or `"deny"` |
231
+ | `denyMessage` | string | No | Deny reason shown to Claude (`deny` only). Default: `"Permission denied by caller"` |
232
+ | `interrupt` | boolean | No | When true, denying also interrupts the whole agent (`deny` only). Default: `false` |
233
+ | `pollOptions` | object | No | Fine-grained poll control options (see below) |
234
+ | `permissionOptions` | object | No | Advanced permission response options (see below) |
235
+
236
+ <details>
237
+ <summary><code>pollOptions</code> object parameters (9 fine-grained poll controls)</summary>
238
+
239
+ | Parameter | Type | Description |
240
+ | ---------------------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------- |
241
+ | `pollOptions.includeTools` | boolean | When true, includes `availableTools` (`poll` only). Default: `false` (omitted until session init is received) |
242
+ | `pollOptions.includeEvents` | boolean | When false, omits `events` (but `nextCursor` still advances). Default: `true` |
243
+ | `pollOptions.includeActions` | boolean | When false, omits `actions[]` even if `waiting_permission`. Default: `true` |
244
+ | `pollOptions.includeResult` | boolean | When false, omits top-level `result` even when `idle`/`error`. Default: `true` |
245
+ | `pollOptions.includeUsage` | boolean | Include `result.usage` (default: true in full mode, false in minimal mode) |
246
+ | `pollOptions.includeModelUsage` | boolean | Include `result.modelUsage` (default: true in full mode, false in minimal mode) |
247
+ | `pollOptions.includeStructuredOutput` | boolean | Include `result.structuredOutput` (default: true in full mode, false in minimal mode) |
248
+ | `pollOptions.includeTerminalEvents` | boolean | When true, keeps terminal `result`/`error` events in `events` even if top-level `result` is included. Default: `false` in `"minimal"`, `true` in `"full"` |
249
+ | `pollOptions.includeProgressEvents` | boolean | When true, includes progress events (`tool_progress`, `auth_status`) in the events stream. Default: `false` in `"minimal"`, `true` in `"full"` |
250
+
251
+ </details>
252
+
253
+ <details>
254
+ <summary><code>permissionOptions</code> object parameters (2 advanced permission response options)</summary>
196
255
 
197
- Enable or disable `bypassPermissions` mode at runtime without restarting the server.
256
+ | Parameter | Type | Description |
257
+ | ---------------------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------- |
258
+ | `permissionOptions.updatedInput` | object | Modified tool input to run (`allow` only). Default: none |
259
+ | `permissionOptions.updatedPermissions` | array | Permission rule updates suggested/applied (`allow` only). Default: none |
198
260
 
199
- | Parameter | Type | Required | Description |
200
- | --------- | ------ | -------- | -------------------------------------------------------- |
201
- | `action` | string | Yes | `"enable_bypass"`, `"disable_bypass"`, or `"get_config"` |
261
+ </details>
202
262
 
203
- **Returns:** `{ allowBypass, message, isError? }`
263
+ **Returns (poll and respond_permission):** `{ sessionId, status, pollInterval?, cursorResetTo?, truncated?, truncatedFields?, events, nextCursor?, availableTools?, actions?, result?, cancelledAt?, cancelledReason?, cancelledSource?, lastEventId?, lastToolUseId? }`
264
+
265
+ Notes:
266
+
267
+ - On error (e.g. invalid arguments, missing/expired session): `{ sessionId, isError: true, error }`
268
+ - Always treat `cursor` as an incremental position: store `nextCursor` and pass it back on the next poll to avoid replaying old events.
269
+ - If `cursorResetTo` is present, your `cursor` was too old (events were evicted); reset your cursor to `cursorResetTo`.
270
+ - For safety, de-duplicate events by `event.id` on the client side.
271
+ - If `truncated=true`, the server intentionally limited the payload (e.g. `maxEvents`) — continue polling with `nextCursor`.
272
+ - In `"minimal"` mode (default): assistant message events are slimmed (strips `usage`, `model`, `id`, `cache_control` from content blocks); noisy progress events (`tool_progress`, `auth_status`) are filtered out; `lastEventId`/`lastToolUseId` are omitted; `AgentResult` omits `durationApiMs`/`sessionTotalTurns`/`sessionTotalCostUsd`. Use `responseMode: "full"` or individual `include*` flags to restore any of these.
204
273
 
205
274
  ## Usage Example
206
275
 
207
276
  ```python
208
- # 1. Start a new session
209
- result = await mcp.call_tool("claude_code", {
277
+ # 1) Start a new session (async start)
278
+ start = await mcp.call_tool("claude_code", {
210
279
  "prompt": "Fix the authentication bug in src/auth.ts",
211
280
  "cwd": "/path/to/project",
212
281
  "allowedTools": ["Read", "Edit", "Bash", "Glob", "Grep"],
213
- "permissionMode": "acceptEdits"
214
- })
215
- session_id = json.loads(result)["sessionId"]
216
-
217
- # 2. Continue the session
218
- result = await mcp.call_tool("claude_code_reply", {
219
- "sessionId": session_id,
220
- "prompt": "Now add unit tests for the fix"
221
- })
222
-
223
- # 3. List all sessions
224
- result = await mcp.call_tool("claude_code_session", {
225
- "action": "list"
226
- })
227
-
228
- # 4. Cancel a running session
229
- result = await mcp.call_tool("claude_code_session", {
230
- "action": "cancel",
231
- "sessionId": session_id
282
+ "advanced": {
283
+ "effort": "high",
284
+ "maxBudgetUsd": 5.0
285
+ }
232
286
  })
287
+ session_id = json.loads(start)["sessionId"]
288
+ cursor = None
289
+
290
+ # 2) Poll until idle/error/cancelled
291
+ while True:
292
+ polled = await mcp.call_tool("claude_code_check", {
293
+ "action": "poll",
294
+ "sessionId": session_id,
295
+ "cursor": cursor,
296
+ "pollOptions": {
297
+ "includeProgressEvents": True
298
+ }
299
+ })
300
+ data = json.loads(polled)
301
+ cursor = data.get("nextCursor", cursor)
302
+
303
+ # If permission is needed, approve/deny via respond_permission
304
+ for action in data.get("actions", []) or []:
305
+ if action.get("type") == "permission":
306
+ await mcp.call_tool("claude_code_check", {
307
+ "action": "respond_permission",
308
+ "sessionId": session_id,
309
+ "requestId": action["requestId"],
310
+ "decision": "allow"
311
+ })
312
+
313
+ # Final result is available when status becomes idle/error
314
+ if data.get("status") in ["idle", "error", "cancelled"]:
315
+ final_result = data.get("result")
316
+ break
317
+
318
+ # 3) Manage sessions (list/get/cancel)
319
+ result = await mcp.call_tool("claude_code_session", {"action": "list"})
233
320
  ```
234
321
 
235
322
  ## Windows Support
@@ -285,25 +372,25 @@ setx CLAUDE_CODE_GIT_BASH_PATH "C:\Program Files\Git\bin\bash.exe"
285
372
 
286
373
  ## Security
287
374
 
288
- - **`permissionMode` defaults to `"dontAsk"`** the agent will deny any operation not pre-approved, avoiding interactive prompts that would hang in MCP context.
289
- - **`bypassPermissions` is disabled by default.** Use the `claude_code_configure` tool with action `enable_bypass` to enable it at runtime.
290
- - **Environment variables are inherited** — the spawned Claude Code process inherits all environment variables (including `ANTHROPIC_API_KEY`) from the parent process by default. The `env` parameter **merges** with `process.env` (user-provided values take precedence), so you can safely add or override individual variables without losing existing ones.
291
- - Use `tools` / `disallowedTools` to restrict the base set of tools the agent can use. Use `allowedTools` to specify which tools are auto-approved without prompting.
292
- - `maxTurns` and `maxBudgetUsd` prevent runaway execution.
375
+ - **Async permission approvals**when a tool call needs approval, the session transitions to `waiting_permission` and surfaces requests via `claude_code_check` (`actions[]`).
376
+ - **No runtime privilege escalation tool** permission decisions are per-session (allow/deny lists + explicit approvals), and the server does not expose a `claude_code_configure` bypass switch.
377
+ - **Environment variables are inherited** — the spawned Claude Code process inherits all environment variables (including `ANTHROPIC_API_KEY`) from the parent process by default. The `advanced.env` parameter **merges** with `process.env` (user-provided values take precedence), so you can safely add or override individual variables without losing existing ones.
378
+ - Tool visibility vs approvals:
379
+ - Use `advanced.tools` to restrict which tools the agent can *see* (hidden tools cannot be called).
380
+ - Use `allowedTools` to auto-approve specific tools without prompting (the SDK may still prompt for path-based restrictions like `blockedPath`).
381
+ - Use `disallowedTools` to hard-block tools; they are denied even if later approved via `claude_code_check`.
382
+ - `maxTurns` and `advanced.maxBudgetUsd` prevent runaway execution.
293
383
  - Sessions auto-expire after 30 minutes of inactivity.
294
384
 
295
385
  ## Environment Variables
296
386
 
297
387
  All environment variables are optional. They are set on the MCP server process (not on the Claude Code child process — for that, use the `env` tool parameter).
298
388
 
299
- | Variable | Description | Default |
300
- | --- | --- | --- |
301
- | `CLAUDE_CODE_GIT_BASH_PATH` | Path to `bash.exe` on Windows (see [Windows Support](#windows-support)) | Auto-detected |
302
- | `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME` | Set to `1` to allow `claude_code_reply` to resume from on-disk transcripts when the in-memory session is missing | `0` (disabled) |
303
- | `CLAUDE_CODE_MCP_ALLOW_SENSITIVE_SESSION_DETAILS` | Set to `1` to allow `claude_code_session` to return `cwd`, `systemPrompt`, `agents`, `additionalDirectories` | `0` (disabled) |
304
- | `CLAUDE_CODE_MCP_SESSION_TTL_MS` | Idle session time-to-live in milliseconds | `1800000` (30 min) |
305
- | `CLAUDE_CODE_MCP_RUNNING_SESSION_MAX_MS` | Maximum wall-clock time for a running session before forced cleanup | `14400000` (4 hr) |
306
- | `CLAUDE_CODE_MCP_CLEANUP_INTERVAL_MS` | How often the cleanup timer runs | `60000` (1 min) |
389
+ | Variable | Description | Default |
390
+ | ------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------- | ------------------ |
391
+ | `CLAUDE_CODE_GIT_BASH_PATH` | Path to `bash.exe` on Windows (see [Windows Support](#windows-support)) | Auto-detected |
392
+ | `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME` | Set to `1` to allow `claude_code_reply` to resume from on-disk transcripts when the in-memory session is missing | `0` (disabled) |
393
+ | `CLAUDE_CODE_MCP_RESUME_SECRET` | HMAC secret used to validate `resumeToken` for disk resume fallback (recommended if disk resume is enabled) | *(unset)* |
307
394
 
308
395
  ### How to configure
309
396
 
@@ -317,7 +404,7 @@ All environment variables are optional. They are set on the MCP server process (
317
404
  "args": ["-y", "@leo000001/claude-code-mcp"],
318
405
  "env": {
319
406
  "CLAUDE_CODE_MCP_ALLOW_DISK_RESUME": "1",
320
- "CLAUDE_CODE_MCP_SESSION_TTL_MS": "3600000"
407
+ "CLAUDE_CODE_MCP_RESUME_SECRET": "change-me"
321
408
  }
322
409
  }
323
410
  }
@@ -333,7 +420,7 @@ args = ["-y", "@leo000001/claude-code-mcp"]
333
420
 
334
421
  [mcp_servers.claude-code.env]
335
422
  CLAUDE_CODE_MCP_ALLOW_DISK_RESUME = "1"
336
- CLAUDE_CODE_MCP_SESSION_TTL_MS = "3600000"
423
+ CLAUDE_CODE_MCP_RESUME_SECRET = "change-me"
337
424
  ```
338
425
 
339
426
  **System-wide** — set via your shell profile or OS settings so all processes inherit them:
@@ -341,9 +428,11 @@ CLAUDE_CODE_MCP_SESSION_TTL_MS = "3600000"
341
428
  ```bash
342
429
  # bash / zsh
343
430
  export CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1
431
+ export CLAUDE_CODE_MCP_RESUME_SECRET=change-me
344
432
 
345
433
  # PowerShell (permanent, requires new terminal)
346
434
  setx CLAUDE_CODE_MCP_ALLOW_DISK_RESUME 1
435
+ setx CLAUDE_CODE_MCP_RESUME_SECRET change-me
347
436
  ```
348
437
 
349
438
  ## Development
@@ -364,13 +453,9 @@ MCP Client ←→ (stdio/JSON-RPC) ←→ MCP Server
364
453
  └── Claude Agent SDK (query())
365
454
  ```
366
455
 
367
- **Session persistence:** The MCP server's Session Manager holds **in-memory** session metadata and a snapshot of session options (e.g. `permissionMode`, tool config, limits, `cwd`). This metadata is **not** persisted to disk by the MCP server. The actual conversation history is persisted to disk by the Claude Code CLI (under `~/.claude/projects/`) — this is managed by the SDK, not by this MCP server. By default, if the MCP server restarts or the session expires from memory, `claude_code_reply` will return `SESSION_NOT_FOUND` even though the CLI transcript may still exist on disk. You can opt into disk-resume behavior by setting `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1`.
368
-
369
- **Session cleanup tuning (optional):** Configure in-memory session cleanup with:
456
+ **Session persistence:** The MCP server's Session Manager holds **in-memory** session metadata, a snapshot of session options (tool config, limits, `cwd`, allow/deny lists, etc.), and an event buffer used by `claude_code_check`. This metadata is **not** persisted to disk by the MCP server. The actual conversation history is persisted to disk by the Claude Code CLI (under `~/.claude/projects/`) — this is managed by the SDK, not by this MCP server. By default, if the MCP server restarts or the session expires from memory, `claude_code_reply` will return `SESSION_NOT_FOUND` even though the CLI transcript may still exist on disk. You can opt into disk-resume behavior by setting `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1`.
370
457
 
371
- - `CLAUDE_CODE_MCP_SESSION_TTL_MS` (default: 1800000)
372
- - `CLAUDE_CODE_MCP_RUNNING_SESSION_MAX_MS` (default: 14400000)
373
- - `CLAUDE_CODE_MCP_CLEANUP_INTERVAL_MS` (default: 60000)
458
+ Sessions are automatically cleaned up after 30 minutes of idle time, or after 4 hours of continuous running.
374
459
 
375
460
  **Turn/Cost semantics:** `numTurns` and `totalCostUsd` are per-call increments. For cumulative per-session totals, use `sessionTotalTurns` and `sessionTotalCostUsd`. When `forkSession=true`, the returned `sessionId` (and `sessionTotal*`) refer to the forked session; the original session totals are preserved.
376
461
 
@@ -382,6 +467,7 @@ MCP server validation/policy errors are returned as `Error [CODE]: message` wher
382
467
  - `SESSION_NOT_FOUND` — session not found in memory (expired or server restarted)
383
468
  - `SESSION_BUSY` — session currently running
384
469
  - `PERMISSION_DENIED` — operation not allowed by server policy
470
+ - `PERMISSION_REQUEST_NOT_FOUND` — permission request ID not found (already finished or expired)
385
471
  - `TIMEOUT` — operation timed out
386
472
  - `CANCELLED` — session was cancelled
387
473
  - `INTERNAL` — unexpected error or protocol mismatch
package/SECURITY.md CHANGED
@@ -18,9 +18,9 @@ We aim to acknowledge reports within 48 hours and provide a fix or mitigation pl
18
18
 
19
19
  ## Security Considerations
20
20
 
21
- - `bypassPermissions` mode is disabled by default and requires explicit opt-in via the `claude_code_configure` tool at runtime
21
+ - This server uses an async permission flow: when a tool call needs approval, the session pauses (`waiting_permission`) and surfaces requests via `claude_code_check` (`actions[]`). Callers must explicitly approve/deny via `respond_permission`.
22
22
  - The MCP server uses the Claude Agent SDK's bundled CLI (`cli.js`), not the system-installed `claude` binary
23
23
  - Session metadata is held in-memory only and is not persisted to disk by the MCP server (the SDK's CLI persists conversation history separately)
24
- - Disk resume is disabled by default. If you set `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1`, the server may resume sessions from the CLI's on-disk transcript even if in-memory metadata is missing.
25
- - `claude_code_session` redacts sensitive fields by default; `includeSensitive` requires `CLAUDE_CODE_MCP_ALLOW_SENSITIVE_SESSION_DETAILS=1`
24
+ - Disk resume is disabled by default (`CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=0`). If you set `CLAUDE_CODE_MCP_ALLOW_DISK_RESUME=1`, disk resume fallback also requires `CLAUDE_CODE_MCP_RESUME_SECRET` (default: unset) and a valid `resumeToken` from `claude_code`/`claude_code_reply`.
25
+ - `claude_code_session` redacts sensitive fields (cwd, systemPrompt, agents, additionalDirectories) by default; use `includeSensitive=true` to include them
26
26
  - Sessions auto-expire after 30 minutes of inactivity