nanocode-cli 0.6.3__tar.gz → 0.6.4__tar.gz

{nanocode_cli-0.6.3/nanocode_cli.egg-info → nanocode_cli-0.6.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nanocode-cli
-Version: 0.6.3
+Version: 0.6.4
 Summary: A small terminal coding agent written in Python
 Author-email: hit9 <hit9@icloud.com>
 License-Expression: BSD-3-Clause
@@ -53,6 +53,7 @@ nanocode is pre-1.0 software. Commands, configuration, and tool behavior may cha
 - **Stale-edit protection**: `line:hash` anchors reject edits when the target code has drifted.
 - **Project-aware navigation**: Use the symbol index to jump through outlines, references, implementors, call chains, and changed files quickly.
 - **Recoverable context**: Tool output stays bounded in the prompt, while raw `tr.N` results remain recallable.
+- **Session recovery**: Resume saved work with `nanocode --resume`, including restored conversation history.
 - **Cache-aware context**: Stable sections stay early and noisy working state stays late to improve prompt-cache reuse.
 - **Focused working memory**: `Note` separates goal, plan, and known facts from noisy execution logs.
 - **MCP integration**: Connect to remote (HTTP) or local (stdio) Model Context Protocol servers and call their tools.
@@ -60,6 +61,8 @@ nanocode is pre-1.0 software. Commands, configuration, and tool behavior may cha
 
 ## Install
 
+Install with uv:
+
 ```sh
 uv tool install nanocode-cli
 ```
@@ -77,9 +80,15 @@ uv sync --extra dev
 uv run nanocode
 ```
 
-## Usage
+## Quick Start
+
+Create a config file:
+
+```sh
+nanocode --init-config
+```
 
-Start the CLI:
+Edit `~/.nanocode/config.toml`, then start:
 
 ```sh
 nanocode
@@ -89,15 +98,38 @@ Useful arguments:
 
 - `--config <path>`: use a TOML config file.
 - `--init-config`: create a default config file.
+- `--resume [UID]`: resume a saved session; without `UID`, resumes `latest`.
 - `--yolo`: skip confirmations for mutating tools.
+- `--mcp <selector>`: choose which configured MCP servers to enable.
+- `--debug`: write model I/O debug traces.
 - `-v`, `--version`: show the version.
 
-During a running turn, the `+>` prompt accepts follow-up input for the next model request.
+During a running turn, type into the `+>` prompt to add follow-up input for the next model request.
+
+## Sessions
+
+nanocode saves recoverable non-empty sessions under `[paths] data_dir` and prints a restore command on exit:
+
+```sh
+Resume with: nanocode --resume <session-id>
+```
+
+Resume by id, or use the latest saved session:
+
+```sh
+nanocode --resume <session-id>
+nanocode --resume latest
+nanocode --resume last
+```
+
+Restored sessions replay the visible conversation history once; tool summaries are shown, raw tool result bodies are not. `/status` shows the active session id. Old session files are cleaned up after `runtime.session_retention_days` days, defaulting to `7`; set it to `0` to disable cleanup.
+
+## CLI
 
-## Commands
+Commands:
 
 - `/help`: show commands and tools.
-- `/status`: show runtime status.
+- `/status`: show runtime status, including the active session id.
 - `/config`: show active config.
 - `/api [auto|chat|anthropic]`: show or set provider API format.
 - `/debug [on|off]`: toggle model I/O debug traces.
@@ -107,13 +139,13 @@ During a running turn, the `+>` prompt accepts follow-up input for the next mode
 - `/provider [NAME]`: show or set provider.
 - `/model [MODEL]`: show or set model.
 - `/reason`: choose reasoning effort.
-- `/set KEY VALUE`: set provider/runtime values.
+- `/set KEY VALUE`: set supported provider/runtime values for the current session.
 - `/yolo`: toggle tool confirmations.
 - `/exit`, `/quit`: exit.
 
 Interactive selectors support `j`/`k`, arrows, `/` search, Enter, and Esc. Input supports history, completion, and `Ctrl-R` history search.
 
-## Tools
+Tools:
 
 - File: `Read`, `LineCount`, `List`, `Find`, `Search`.
 - Code index: `InspectCode`.
@@ -121,30 +153,30 @@ Interactive selectors support `j`/`k`, arrows, `/` search, Enter, and Esc. Input
 - Shell: `Bash`, `Git`.
 - Tool results: `Recall`.
 - Working notes: `Note`.
-- Ask the user: `Question` asks one or more questions (optional choices, previews, recommended) when intent is genuinely ambiguous.
-- MCP: `MCP` calls tools on configured MCP servers.
+- Ask the user: `Question`.
+- MCP: `MCP`.
 
 `Read`, `Search`, and `InspectCode` return line anchors where useful. `Edit` uses current `line:hash` anchors to reject stale edits.
 
 ## Configuration
 
-Run:
+Default config location:
 
-```sh
-nanocode --init-config
+```text
+~/.nanocode/config.toml
 ```
 
-Default config location is `~/.nanocode/config.toml`.
-
 Main fields:
 
 - `[provider] active = "name"`
 - `[provider.<name>]`: `url`, `key`, `model`, `api`, `prompt_cache_key`, `available_models`, `reasoning`, `chat_reasoning`, `temperature`, `timeout`
 - `[paths] data_dir`
-- `[runtime] shell_timeout`, `max_agent_steps`, `max_context_tokens`, `yolo`
+- `[runtime] shell_timeout`, `max_agent_steps`, `max_context_tokens`, `check_updates`, `update_check_interval_hours`, `session_retention_days`, `yolo`, `debug`
 
 `api = "auto"` chooses between Chat Completions and Anthropic Messages using provider/model profiles. `prompt_cache_key = "auto"` derives a stable key from provider, model, workspace, and tool schema names.
 
+Runtime flags such as `--yolo`, `--debug`, and `--mcp` apply to resumed sessions too. Saved sessions do not carry their old runtime config forward.
+
 ## MCP
 
 nanocode connects to [Model Context Protocol](https://modelcontextprotocol.io) servers and exposes their tools through the `MCP` tool. Configure each server under `[mcp.<name>]`. A server is either `url` (remote) or `command` (local), never both.
@@ -163,7 +195,7 @@ auth = "oauth"                              # browser login via /mcp login <serv
 enabled = true
 ```
 
-Local server over stdio (launched as a subprocess):
+Local server over stdio:
 
 ```toml
 [mcp.filesystem]
@@ -182,16 +214,16 @@ Manage servers at runtime:
 - `/mcp refresh [server]`: rediscover servers.
 - `/mcp login <server>` / `/mcp logout <server>`: OAuth login and logout.
 
-## Tested Providers
+## Providers
 
 The following providers have been tested with nanocode:
 
 - **deepseek**: DeepSeek API
 - **opencode**: OpenCode API
-- **aliyun**: Alibaba Cloud (Tongyi Qianwen) API via Chat Completions
+- **aliyun**: Alibaba Cloud Tongyi Qianwen API via Chat Completions
 - **llama.cpp**: Local inference via llama.cpp server
 
-## Context Design
+## Context Model
 
 Each model request is built manually from explicit messages. Stable context comes first, conversation stays as messages, working memory follows, and the latest file state is appended at the end.
 
@@ -202,16 +234,16 @@ model request
 |   concise agent contract and tool rules          |
 +--------------------------------------------------+
 | user                                             |
-|   Environment                                   |
+|   Environment                                    |
 +--------------------------------------------------+
-| user/assistant                                  |
-|   conversation, compacted summaries, tools      |
+| user/assistant                                   |
+|   conversation, compacted summaries, tools       |
 +--------------------------------------------------+
 | user                                             |
-|   Memory: goal, plan, known, date               |
+|   Memory: goal, plan, known, date                |
 +--------------------------------------------------+
 | user                                             |
-|   FILE STATE: latest Read/Edit file view        |
+|   FILE STATE: latest Read/Edit file view         |
 +--------------------------------------------------+
 ```
 

{nanocode_cli-0.6.3 → nanocode_cli-0.6.4}/README.md

@@ -17,6 +17,7 @@ nanocode is pre-1.0 software. Commands, configuration, and tool behavior may cha
 - **Stale-edit protection**: `line:hash` anchors reject edits when the target code has drifted.
 - **Project-aware navigation**: Use the symbol index to jump through outlines, references, implementors, call chains, and changed files quickly.
 - **Recoverable context**: Tool output stays bounded in the prompt, while raw `tr.N` results remain recallable.
+- **Session recovery**: Resume saved work with `nanocode --resume`, including restored conversation history.
 - **Cache-aware context**: Stable sections stay early and noisy working state stays late to improve prompt-cache reuse.
 - **Focused working memory**: `Note` separates goal, plan, and known facts from noisy execution logs.
 - **MCP integration**: Connect to remote (HTTP) or local (stdio) Model Context Protocol servers and call their tools.
@@ -24,6 +25,8 @@ nanocode is pre-1.0 software. Commands, configuration, and tool behavior may cha
 
 ## Install
 
+Install with uv:
+
 ```sh
 uv tool install nanocode-cli
 ```
@@ -41,9 +44,15 @@ uv sync --extra dev
 uv run nanocode
 ```
 
-## Usage
+## Quick Start
+
+Create a config file:
+
+```sh
+nanocode --init-config
+```
 
-Start the CLI:
+Edit `~/.nanocode/config.toml`, then start:
 
 ```sh
 nanocode
@@ -53,15 +62,38 @@ Useful arguments:
 
 - `--config <path>`: use a TOML config file.
 - `--init-config`: create a default config file.
+- `--resume [UID]`: resume a saved session; without `UID`, resumes `latest`.
 - `--yolo`: skip confirmations for mutating tools.
+- `--mcp <selector>`: choose which configured MCP servers to enable.
+- `--debug`: write model I/O debug traces.
 - `-v`, `--version`: show the version.
 
-During a running turn, the `+>` prompt accepts follow-up input for the next model request.
+During a running turn, type into the `+>` prompt to add follow-up input for the next model request.
+
+## Sessions
+
+nanocode saves recoverable non-empty sessions under `[paths] data_dir` and prints a restore command on exit:
+
+```sh
+Resume with: nanocode --resume <session-id>
+```
+
+Resume by id, or use the latest saved session:
+
+```sh
+nanocode --resume <session-id>
+nanocode --resume latest
+nanocode --resume last
+```
+
+Restored sessions replay the visible conversation history once; tool summaries are shown, raw tool result bodies are not. `/status` shows the active session id. Old session files are cleaned up after `runtime.session_retention_days` days, defaulting to `7`; set it to `0` to disable cleanup.
+
+## CLI
 
-## Commands
+Commands:
 
 - `/help`: show commands and tools.
-- `/status`: show runtime status.
+- `/status`: show runtime status, including the active session id.
 - `/config`: show active config.
 - `/api [auto|chat|anthropic]`: show or set provider API format.
 - `/debug [on|off]`: toggle model I/O debug traces.
@@ -71,13 +103,13 @@ During a running turn, the `+>` prompt accepts follow-up input for the next mode
 - `/provider [NAME]`: show or set provider.
 - `/model [MODEL]`: show or set model.
 - `/reason`: choose reasoning effort.
-- `/set KEY VALUE`: set provider/runtime values.
+- `/set KEY VALUE`: set supported provider/runtime values for the current session.
 - `/yolo`: toggle tool confirmations.
 - `/exit`, `/quit`: exit.
 
 Interactive selectors support `j`/`k`, arrows, `/` search, Enter, and Esc. Input supports history, completion, and `Ctrl-R` history search.
 
-## Tools
+Tools:
 
 - File: `Read`, `LineCount`, `List`, `Find`, `Search`.
 - Code index: `InspectCode`.
@@ -85,30 +117,30 @@ Interactive selectors support `j`/`k`, arrows, `/` search, Enter, and Esc. Input
 - Shell: `Bash`, `Git`.
 - Tool results: `Recall`.
 - Working notes: `Note`.
-- Ask the user: `Question` asks one or more questions (optional choices, previews, recommended) when intent is genuinely ambiguous.
-- MCP: `MCP` calls tools on configured MCP servers.
+- Ask the user: `Question`.
+- MCP: `MCP`.
 
 `Read`, `Search`, and `InspectCode` return line anchors where useful. `Edit` uses current `line:hash` anchors to reject stale edits.
 
 ## Configuration
 
-Run:
+Default config location:
 
-```sh
-nanocode --init-config
+```text
+~/.nanocode/config.toml
 ```
 
-Default config location is `~/.nanocode/config.toml`.
-
 Main fields:
 
 - `[provider] active = "name"`
 - `[provider.<name>]`: `url`, `key`, `model`, `api`, `prompt_cache_key`, `available_models`, `reasoning`, `chat_reasoning`, `temperature`, `timeout`
 - `[paths] data_dir`
-- `[runtime] shell_timeout`, `max_agent_steps`, `max_context_tokens`, `yolo`
+- `[runtime] shell_timeout`, `max_agent_steps`, `max_context_tokens`, `check_updates`, `update_check_interval_hours`, `session_retention_days`, `yolo`, `debug`
 
 `api = "auto"` chooses between Chat Completions and Anthropic Messages using provider/model profiles. `prompt_cache_key = "auto"` derives a stable key from provider, model, workspace, and tool schema names.
 
+Runtime flags such as `--yolo`, `--debug`, and `--mcp` apply to resumed sessions too. Saved sessions do not carry their old runtime config forward.
+
 ## MCP
 
 nanocode connects to [Model Context Protocol](https://modelcontextprotocol.io) servers and exposes their tools through the `MCP` tool. Configure each server under `[mcp.<name>]`. A server is either `url` (remote) or `command` (local), never both.
@@ -127,7 +159,7 @@ auth = "oauth"                              # browser login via /mcp login <serv
 enabled = true
 ```
 
-Local server over stdio (launched as a subprocess):
+Local server over stdio:
 
 ```toml
 [mcp.filesystem]
@@ -146,16 +178,16 @@ Manage servers at runtime:
 - `/mcp refresh [server]`: rediscover servers.
 - `/mcp login <server>` / `/mcp logout <server>`: OAuth login and logout.
 
-## Tested Providers
+## Providers
 
 The following providers have been tested with nanocode:
 
 - **deepseek**: DeepSeek API
 - **opencode**: OpenCode API
-- **aliyun**: Alibaba Cloud (Tongyi Qianwen) API via Chat Completions
+- **aliyun**: Alibaba Cloud Tongyi Qianwen API via Chat Completions
 - **llama.cpp**: Local inference via llama.cpp server
 
-## Context Design
+## Context Model
 
 Each model request is built manually from explicit messages. Stable context comes first, conversation stays as messages, working memory follows, and the latest file state is appended at the end.
 
@@ -166,16 +198,16 @@ model request
 |   concise agent contract and tool rules          |
 +--------------------------------------------------+
 | user                                             |
-|   Environment                                   |
+|   Environment                                    |
 +--------------------------------------------------+
-| user/assistant                                  |
-|   conversation, compacted summaries, tools      |
+| user/assistant                                   |
+|   conversation, compacted summaries, tools       |
 +--------------------------------------------------+
 | user                                             |
-|   Memory: goal, plan, known, date               |
+|   Memory: goal, plan, known, date                |
 +--------------------------------------------------+
 | user                                             |
-|   FILE STATE: latest Read/Edit file view        |
+|   FILE STATE: latest Read/Edit file view         |
 +--------------------------------------------------+
 ```
 

{nanocode_cli-0.6.3 → nanocode_cli-0.6.4}/README.zh-CN.md

@@ -15,8 +15,9 @@ nanocode 仍是 1.0 前的软件。稳定版发布前，命令、配置和工具
 - **实时回合控制**：代理还在工作时，也可以追加输入，不打断当前工具流程。
 - **文件状态大脑**：`Read` 和 `Edit` 会构建当前重要文件的带行号视图。
 - **过期编辑保护**：`line:hash` 锚点会在目标代码漂移后拒绝错误编辑。
-- **项目级导航**：通过符号索引快速查看 outline、references 和变更文件。
+- **项目级导航**：通过符号索引快速查看 outline、references、implementors、call chains 和变更文件。
 - **可恢复上下文**：prompt 中的工具输出保持有界，原始 `tr.N` 结果仍可按需召回。
+- **Session 恢复**：通过 `nanocode --resume` 恢复已保存工作，包括重新展示的会话历史。
 - **缓存友好上下文**：稳定内容靠前，嘈杂的工作状态靠后，提高 prompt cache 复用率。
 - **聚焦工作记忆**：`Note` 把 goal、plan、known facts 从嘈杂执行日志中拆出来。
 - **MCP 集成**：连接远程（HTTP）或本地（stdio）的 Model Context Protocol 服务器并调用其工具。
@@ -24,6 +25,8 @@ nanocode 仍是 1.0 前的软件。稳定版发布前，命令、配置和工具
 
 ## 安装
 
+使用 uv 安装：
+
 ```sh
 uv tool install nanocode-cli
 ```
@@ -41,9 +44,15 @@ uv sync --extra dev
 uv run nanocode
 ```
 
-## 使用
+## 快速开始
+
+创建配置文件：
+
+```sh
+nanocode --init-config
+```
 
-启动 CLI：
+编辑 `~/.nanocode/config.toml`，然后启动：
 
 ```sh
 nanocode
@@ -53,15 +62,38 @@ nanocode
 
 - `--config <path>`：使用指定 TOML 配置文件。
 - `--init-config`：创建默认配置文件。
+- `--resume [UID]`：恢复已保存的 session；不传 `UID` 时恢复 `latest`。
 - `--yolo`：跳过会修改环境的工具确认。
+- `--mcp <selector>`：选择启用哪些已配置的 MCP 服务器。
+- `--debug`：写入模型 I/O debug trace。
 - `-v`, `--version`：显示版本。
 
-代理运行中，`+>` 提示符可以接收追加输入，并在下一次模型请求中发送。
+代理运行中，可以在 `+>` 提示符里输入追加内容，发送到下一次模型请求。
 
-## 命令
+## Sessions
+
+nanocode 会把有可恢复内容的非空 session 保存到 `[paths] data_dir` 下，并在退出时打印恢复命令：
+
+```sh
+Resume with: nanocode --resume <session-id>
+```
+
+按 id 恢复，或恢复最近保存的 session：
+
+```sh
+nanocode --resume <session-id>
+nanocode --resume latest
+nanocode --resume last
+```
+
+恢复后会在启动时重放一次可见会话历史；工具摘要会显示，但不会打印原始 tool result 正文。`/status` 会显示当前 session id。早于 `runtime.session_retention_days` 的 session 文件会在启动时清理，默认 `7` 天；设置为 `0` 可关闭清理。
+
+## CLI
+
+命令：
 
 - `/help`：显示命令和工具。
-- `/status`：显示运行状态。
+- `/status`：显示运行状态，包括当前 session id。
 - `/config`：显示当前配置。
 - `/api [auto|chat|anthropic]`：显示或设置 provider API 格式。
 - `/debug [on|off]`：切换模型 I/O debug trace。
@@ -71,13 +103,13 @@ nanocode
 - `/provider [NAME]`：显示或设置 provider。
 - `/model [MODEL]`：显示或设置模型。
 - `/reason`：选择 reasoning effort。
-- `/set KEY VALUE`：设置 provider/runtime 值。
+- `/set KEY VALUE`：设置当前 session 支持的 provider/runtime 值。
 - `/yolo`：切换工具确认。
 - `/exit`, `/quit`：退出。
 
 交互选择器支持 `j`/`k`、方向键、`/` 搜索、Enter 和 Esc。输入框支持历史、补全和 `Ctrl-R` 历史搜索。
 
-## 工具
+工具：
 
 - 文件：`Read`, `LineCount`, `List`, `Find`, `Search`。
 - 代码索引：`InspectCode`。
@@ -85,30 +117,30 @@ nanocode
 - Shell：`Bash`, `Git`。
 - 工具结果：`Recall`。
 - 工作笔记：`Note`。
-- 询问用户：`Question` 在意图真正歧义时向用户提一个或多个问题（可选 choices、previews、recommended）。
-- MCP：`MCP` 调用已配置 MCP 服务器上的工具。
+- 询问用户：`Question`。
+- MCP：`MCP`。
 
 `Read`、`Search` 和 `InspectCode` 会在合适时返回行锚点。`Edit` 使用当前 `line:hash` 锚点拒绝过期编辑。
 
 ## 配置
 
-运行：
+默认配置位置：
 
-```sh
-nanocode --init-config
+```text
+~/.nanocode/config.toml
 ```
 
-默认配置位置是 `~/.nanocode/config.toml`。
-
 主要字段：
 
 - `[provider] active = "name"`
 - `[provider.<name>]`：`url`, `key`, `model`, `api`, `prompt_cache_key`, `available_models`, `reasoning`, `chat_reasoning`, `temperature`, `timeout`
 - `[paths] data_dir`
-- `[runtime] shell_timeout`, `max_agent_steps`, `max_context_tokens`, `yolo`
+- `[runtime] shell_timeout`, `max_agent_steps`, `max_context_tokens`, `check_updates`, `update_check_interval_hours`, `session_retention_days`, `yolo`, `debug`
 
 `api = "auto"` 会根据 provider/model profile 在 Chat Completions 和 Anthropic Messages 之间选择。`prompt_cache_key = "auto"` 会根据 provider、model、workspace 和工具 schema 名称生成稳定 key。
 
+`--yolo`、`--debug` 和 `--mcp` 等 runtime flags 对恢复的 session 同样生效。保存的 session 不会携带旧 runtime config。
+
 ## MCP
 
 nanocode 可连接 [Model Context Protocol](https://modelcontextprotocol.io) 服务器，并通过 `MCP` 工具暴露其工具。每个服务器配置在 `[mcp.<name>]` 下，且只能二选一：`url`（远程）或 `command`（本地）。
@@ -127,7 +159,7 @@ auth = "oauth"                              # 通过 /mcp login <server> 在浏
 enabled = true
 ```
 
-通过 stdio 的本地服务器（作为子进程启动）：
+通过 stdio 的本地服务器：
 
 ```toml
 [mcp.filesystem]
@@ -146,15 +178,16 @@ HTTP 鉴权选项（`auth`、`bearer_token_env_var`、`env_http_headers`）只
 - `/mcp refresh [server]`：重新发现服务器。
 - `/mcp login <server>` / `/mcp logout <server>`：OAuth 登录和登出。
 
-## 已测试的 Provider
+## Providers
 
 以下 provider 已在 nanocode 中测试通过：
 
-- **deepseek**: DeepSeek API
-- **opencode**: OpenCode API
-- **aliyun**: 阿里云通义千问 API（Chat Completions）
-- **llama.cpp**: 通过 llama.cpp 服务端本地推理
-## 上下文设计
+- **deepseek**：DeepSeek API
+- **opencode**：OpenCode API
+- **aliyun**：阿里云通义千问 API（Chat Completions）
+- **llama.cpp**：通过 llama.cpp 服务端本地推理
+
+## 上下文模型
 
 每次模型请求都由 nanocode 手动构建成明确的 messages。稳定上下文在前，会话作为 messages 保留，工作记忆随后，最新文件状态放在末尾。
 
@@ -165,16 +198,16 @@ model request
 |   concise agent contract and tool rules          |
 +--------------------------------------------------+
 | user                                             |
-|   Environment                                   |
+|   Environment                                    |
 +--------------------------------------------------+
-| user/assistant                                  |
-|   conversation, compacted summaries, tools      |
+| user/assistant                                   |
+|   conversation, compacted summaries, tools       |
 +--------------------------------------------------+
 | user                                             |
-|   Memory: goal, plan, known, date               |
+|   Memory: goal, plan, known, date                |
 +--------------------------------------------------+
 | user                                             |
-|   FILE STATE: latest Read/Edit file view        |
+|   FILE STATE: latest Read/Edit file view         |
 +--------------------------------------------------+
 ```