longer-agent 0.1.0 → 0.1.1

package/README.md

@@ -2,7 +2,7 @@
   <img src="https://raw.githubusercontent.com/FelixRuiGao/LongerAgent/main/assets/logo.png" alt="LongerAgent" width="360" />
 </p>
 <p align="center">
-  <strong>A terminal AI coding agent built for long sessions.</strong>
+  <strong>Built to work longer.</strong>
 </p>
 <p align="center">
   English | <a href="./README.zh-CN.md">中文</a>
@@ -14,15 +14,22 @@
   <img alt="Author" src="https://img.shields.io/badge/author-Felix%20Rui%20Gao-4b4bf0?style=flat-square" />
 </p>
 
-Most AI agents crash, loop, or silently lose context when conversations get long. LongerAgent is built from the ground up for sessions that last longer — with a structured log architecture, three-layer context management, and persistent memory that survives across sessions.
+A terminal AI coding agent that manages its own context proactively, runs parallel sub-agents, and lets you message it while it's working.
 
 ![LongerAgent Terminal UI](https://raw.githubusercontent.com/FelixRuiGao/LongerAgent/main/assets/screenshot.png)
 
-https://github.com/user-attachments/assets/e405916d-f205-4995-b4b3-ce81b19af618
+> **Platform:** macOS.
 
-> **Platform:** macOS. Windows is not tested.
->
-> **Security:** LongerAgent does not sandbox commands or require approval before file edits and shell execution. Use it in trusted environments and review what it does.
+## Why LongerAgent
+
+Most coding agents work well for short bursts, then degrade as the session gets longer. LongerAgent is built for the opposite case:
+
+- **Long-running sessions** — context is monitored, summarized, and compacted before it collapses
+- **Interruptible work** — you can send a new message while the agent is still executing
+- **Parallel execution** — delegate exploration and implementation to sub-agents in the same session
+- **Project memory** — `AGENTS.md` and the Important Log survive across sessions and compactions
+
+If you want a terminal-native agent that can stay productive through a long refactor, investigation, or multi-step task, this is the point of the project.
 
 ## Quick Start
 
@@ -37,16 +44,57 @@ longeragent init
 longeragent
 ```
 
+The setup wizard walks you through provider and model selection. If you prefer to edit config yourself, use [configExample.yaml](./configExample.yaml) as the reference.
+
+### First 5 Minutes
+
+Typical first session:
+
+```text
+$ longeragent
+You: audit the auth flow and find dead code
+You: while you do that, also check whether password reset uses the same token path
+You: /summarize keep the final findings and file references
+```
+
+Useful early commands:
+
+```text
+/model       # switch model/provider for the current session
+/thinking    # raise or lower reasoning depth
+/skills      # enable or disable installed skills
+/resume      # reopen an older session from log
+```
+
+> **Safety:** LongerAgent does not sandbox shell commands or file edits. Run it in trusted environments and review what it does.
+
+## Demo
+
+Parallel sub-agents investigating a codebase, an async message mid-task, and context summarization — all in one session.
+
+https://github.com/user-attachments/assets/377fe648-d43c-45da-b111-9434b2a0dc61
+
 ---
 
 ## Highlights
 
-- **Three-layer context management** — sessions that last longer and longer
-- **Parallel sub-agents** — spawn workers for concurrent tasks from YAML call files
-- **Skills system** — install, manage, and create reusable skill packages by itself
+- **Three-layer context management** — hinting, surgical summarization, and full compaction
+- **Parallel sub-agents** — spawn workers from chat or YAML call files
+- **Skills system** — install, manage, and create reusable skill packages from inside the agent
 - **Persistent memory** — `AGENTS.md` files and Important Log survive across sessions and compactions
 - **Async messaging** — talk to the agent while it's mid-task
-- **7 provider families** — Anthropic, OpenAI, Kimi, MiniMax, GLM, OpenRouter, and any OpenRouter-compatible model
+- **7 provider families** — Anthropic, OpenAI, Kimi, MiniMax, GLM, OpenRouter, and OpenRouter-compatible models
+
+## What It Feels Like
+
+LongerAgent is optimized for a specific workflow:
+
+1. Start a real task, not a toy prompt.
+2. Let the agent explore, edit, and test for a while.
+3. Interrupt it with clarifications or side requests without losing momentum.
+4. Keep the session alive by summarizing or compacting instead of restarting from scratch.
+
+That combination is the core product, more than any individual slash command or tool.
 
 ## Usage
 
@@ -222,6 +270,10 @@ pnpm test           # Run tests (vitest)
 pnpm typecheck      # Type check
 ```
 
+## Security
+
+LongerAgent does not sandbox commands or require approval before file edits and shell execution. Use it in trusted environments and review what it does.
+
 ## License
 
 [MIT](./LICENSE)

package/README.zh-CN.md

@@ -2,7 +2,7 @@
   <img src="https://raw.githubusercontent.com/FelixRuiGao/LongerAgent/main/assets/logo.png" alt="LongerAgent" width="360" />
 </p>
 <p align="center">
-  <strong>为长时间会话而生的终端 AI 编程助手。</strong>
+  <strong>Built to work longer.</strong>
 </p>
 <p align="center">
   <a href="./README.md">English</a> | 中文
@@ -14,15 +14,22 @@
   <img alt="Author" src="https://img.shields.io/badge/author-Felix%20Rui%20Gao-4b4bf0?style=flat-square" />
 </p>
 
-大多数 AI 助手在长对话中会崩溃、循环或悄悄丢失上下文。LongerAgent 从底层为长时间会话而设计——结构化日志架构、三层上下文管理、跨会话持久记忆。
+一个能主动管理自身上下文、运行并行子 Agent、并允许你在工作中随时发送消息的终端 AI 编程助手。
 
 ![LongerAgent Terminal UI](https://raw.githubusercontent.com/FelixRuiGao/LongerAgent/main/assets/screenshot.png)
 
-https://github.com/user-attachments/assets/e405916d-f205-4995-b4b3-ce81b19af618
+> **平台：** macOS。
 
-> **平台：** macOS。Windows 未经测试。
->
-> **安全：** LongerAgent 不对命令做沙盒隔离，也不会在文件编辑和 Shell 执行前要求审批。请在可信环境中使用，并留意它的操作。
+## 为什么是 LongerAgent
+
+很多 Coding Agent 在短任务里表现不错，但会话一长就开始失速。LongerAgent 是反过来设计的：
+
+- **长会话** —— 在上下文崩掉之前就监控、压缩、重置
+- **可打断执行** —— Agent 工作到一半时，你还能继续发消息
+- **并行执行** —— 同一会话里把探索和执行拆给多个子 Agent
+- **项目记忆** —— `AGENTS.md` 和 Important Log 跨会话、跨压缩保留
+
+如果你想要的是一个真正适合长时间重构、排查、迭代执行的终端 Agent，这就是它的设计目标。
 
 ## 快速开始
 
@@ -37,16 +44,57 @@ longeragent init
 longeragent
 ```
 
+初始化向导会带你完成 provider 和 model 的选择。如果你更想手动配置，可以直接参考 [configExample.yaml](./configExample.yaml)。
+
+### 前 5 分钟
+
+一个典型的首次会话会是这样：
+
+```text
+$ longeragent
+你：检查一下认证流程，顺便找出死代码
+你：另外也看看 password reset 是否复用了同一条 token 路径
+你：/summarize 保留最终结论和文件引用
+```
+
+前期最常用的几个命令：
+
+```text
+/model       # 切换当前会话使用的模型/provider
+/thinking    # 调高或调低推理深度
+/skills      # 启用或禁用已安装技能
+/resume      # 从日志恢复旧会话
+```
+
+> **安全提示：** LongerAgent 不会对 shell 命令或文件编辑做沙箱隔离。请只在可信环境中使用，并留意它的实际操作。
+
+## 演示
+
+并行子 Agent 调查代码架构、运行中发送异步消息、上下文压缩——一个会话内完成。
+
+https://github.com/user-attachments/assets/377fe648-d43c-45da-b111-9434b2a0dc61
+
 ---
 
 ## 亮点
 
-- **三层上下文管理** —— 会话越来越长，也不怕
-- **并行子 Agent** —— 通过 YAML 调用文件 spawn 多个 worker 并发执行
-- **Skills 系统** —— 自主搜索、安装和管理可复用的技能包
+- **三层上下文管理** —— 提示、精确压缩、全量 compact 三层协作
+- **并行子 Agent** —— 可在对话里或通过 YAML 调用文件派发任务
+- **Skills 系统** —— 直接在 Agent 内安装、管理、创建可复用技能包
 - **持久记忆** —— `AGENTS.md` 和 Important Log 跨会话、跨压缩存续
 - **异步消息** —— Agent 工作中随时发送消息，不用等它完成
-- **7 大 Provider** —— Anthropic、OpenAI、Kimi、MiniMax、GLM、OpenRouter，以及任何 OpenRouter 兼容模型
+- **7 大 Provider** —— Anthropic、OpenAI、Kimi、MiniMax、GLM、OpenRouter，以及 OpenRouter 兼容模型
+
+## 实际使用感受
+
+LongerAgent 重点优化的是这样一条工作流：
+
+1. 直接开始做真实任务，而不是演示型 prompt。
+2. 让 Agent 连续探索、修改、测试一段时间。
+3. 中途随时插话补充要求，不需要重新开局。
+4. 通过 summarize 或 compact 维持会话寿命，而不是上下文一满就从头再来。
+
+这个组合本身，才是 LongerAgent 的核心价值，而不只是某一个命令或工具。
 
 ## 使用
 
@@ -222,6 +270,10 @@ pnpm test           # 运行测试（vitest）
 pnpm typecheck      # 类型检查
 ```
 
+## 安全
+
+LongerAgent 不对命令做沙盒隔离，也不会在文件编辑和 Shell 执行前要求审批。请在可信环境中使用，并留意它的操作。
+
 ## 许可证
 
 [MIT](./LICENSE)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "longer-agent",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "A terminal AI coding agent built for long sessions — three-layer context management, persistent memory, and parallel sub-agents",
   "type": "module",
   "license": "MIT",