Release v6.2.1 inherited ccb config skill

Author: SeemSeam

CHANGELOG.md

@@ -2,6 +2,16 @@
 
 ## Unreleased
 
+## v6.2.1 (2026-05-18)
+
+### Inherited CCB Config Skill Release
+
+- **Inherited `ccb_config` Skill Added**: Claude and Codex installs now inherit a `ccb_config` skill for designing `.ccb/ccb.config`, choosing roles/providers/worktree layout, and updating shared plus per-agent CCB memory.
+- **Inherited Skill Layout Consolidated**: CCB-owned inherited skills now live under `inherit_skills/`; optional `useful_tools/` remain user-installable tools and are not inherited by default.
+- **Ask Guidance Simplified**: injected ask reply guidance is shorter, English-only in source text, skips nested-routing instructions in every ask body, and recognizes more explicit-output requests.
+- **Project Memory Wording Simplified**: generated project/runtime memory now uses shorter submit-once guidance while keeping callback and silence routing rules available where they belong.
+- **Config Memory Routing Clarified**: `ccb_config` memory patterns prefer direct owner-to-next-owner callback handoffs and separate root work packages for parallel chains, without claiming single-task multi-callback fan-in.
+
 ## v6.2.0 (2026-05-17)
 
 ### Callback Ask Chain Release

README.md

@@ -7,7 +7,7 @@
   <img src="https://img.shields.io/badge/Every_Model_Controllable-CF1322?style=for-the-badge" alt="Every Model Controllable">
 </p>
 
-[![Version](https://img.shields.io/badge/version-6.2.0-orange.svg)]()
+[![Version](https://img.shields.io/badge/version-6.2.1-orange.svg)]()
 [![Platform](https://img.shields.io/badge/platform-Linux%20%7C%20macOS%20%7C%20Windows-lightgrey.svg)]()
 
 **English** | [Chinese](README_zh.md)
@@ -74,10 +74,10 @@ Build project-local teams with roles, pane layout, provider state, worktree isol
 <details>
 <summary><b>Latest release highlights</b></summary>
 
-- **Callback ask chains are supported**: active agents can use `ccb ask --callback <agent>` when a child result is needed before replying to the original caller.
-- **Nested ask routing is explicit**: plain nested `ask` from an active CCB task is rejected; use `--callback` for needed results or `--silence` for independent no-result-needed work.
-- **Callback routing is durable**: CCB records callback edges, resumes the parent as a continuation task, and repairs crash windows before submitting missed continuations.
-- **Ask skills explain the workflow**: Claude, Codex, and Droid ask skills now document callback delegation and stop-after-submit behavior.
+- **Inherited `ccb_config` skill added**: Claude and Codex installs now include a CCB-owned skill for designing `.ccb/ccb.config` and team memory.
+- **Inherited skills moved under `inherit_skills/`**: CCB-owned skills are inherited automatically, while optional `useful_tools/` remain user-installable.
+- **Ask guidance is shorter**: CCB injects concise English reply guidance, avoids repeating nested-routing instructions in every ask body, and recognizes more explicit-output requests.
+- **Memory routing is clearer**: `ccb_config` role memory favors direct owner-to-next-owner handoffs and separate root work packages for parallel chains.
 
 See [Release Notes](#release-notes) for the full history.
 
@@ -103,6 +103,25 @@ Tmux copy/paste: drag with the left mouse button to copy, and use `Ctrl+Shift+V`
 
 `.ccb/ccb_memory.md` is the project-wide shared memory document.
 
+<details>
+<summary><b>Config Design Skill</b></summary>
+
+Use `ccb_config` when you want an agent to design or update the CCB team instead of editing config by hand. It is inherited by Claude and Codex installs and focuses on three user-authored files:
+
+- `.ccb/ccb.config` for the team, provider choices, pane layout, and worktree policy
+- `.ccb/ccb_memory.md` for shared project workflow guidance
+- `.ccb/agents/<agent>/memory.md` for per-agent role guidance
+
+Invoke it from a supported provider skill surface, for example:
+
+```text
+$ccb_config Design a team for a Python library with one coordinator, two worktree implementation agents, and one reviewer.
+```
+
+The skill helps choose agent names, providers, `inplace` versus `git-worktree`, compact layout syntax, and whether role instructions belong in shared or per-agent memory. It validates that `.ccb/ccb.config` is the active authority and tells you to restart CCB after file changes are complete.
+
+</details>
+
 <details>
 <summary><b>Layout</b></summary>
 
@@ -186,7 +205,7 @@ CCB v6 currently supports `ccb update` on Linux, macOS, and WSL. A major upgrade
 If you installed from a git checkout with `./install.sh install`, that install now runs in source dev mode:
 
 - Global `ccb` and `ask` link back to the checkout instead of using a copied snapshot
-- CCB-owned skills and helper scripts also follow the live source tree
+- CCB-owned inherited skills under `inherit_skills/` and helper scripts also follow the live source tree
 - Source installs do not participate in startup auto-update prompts
 - Stay on the source/dev track with `git pull` or by switching commits, then rerun `./install.sh install`
 - Or run `ccb update` to install the latest stable release and repoint global `ccb` links to the managed release install
@@ -320,6 +339,16 @@ Thanks to the [Linux.do community](https://linux.do) for testing, feedback, and
 Historical note: older release notes below may mention `askd`, legacy flags, or removed commands. Those references are kept only as changelog history and do not redefine the current CLI surface.
 
 <details open>
+<summary><b>v6.2.1</b> - Inherited CCB Config Skill Release</summary>
+
+- Adds inherited Claude and Codex `ccb_config` skills for designing `.ccb/ccb.config`, choosing agent roles/providers/worktree layout, and updating shared plus per-agent memory.
+- Moves CCB-owned inherited skills under `inherit_skills/` while keeping optional `useful_tools/` user-installable rather than inherited.
+- Shortens injected ask reply guidance, removes nested-routing text from every ask body, keeps injected source text English-only, and expands explicit-output detection.
+- Simplifies project/runtime memory wording and updates `ccb_config` memory-routing examples for direct callback handoffs and separate root work packages.
+
+</details>
+
+<details>
 <summary><b>v6.2.0</b> - Callback Ask Chain Release</summary>
 
 - Adds `ccb ask --callback <agent>` so active agents can delegate work and receive the child result later as a continuation task.
@@ -981,7 +1010,7 @@ See [CHANGELOG.md](CHANGELOG.md) for full details.
 
 - **Zombie Cleanup**: `ccb kill -f` now cleans up orphaned tmux sessions globally (sessions whose parent process has exited)
 - **Mounted Skill**: Optimized to use `pgrep` for daemon detection (~4x faster), extracted to standalone `ccb-mounted` script
-- **Droid Skills**: Added full skill set (cask/gask/lask/oask + ping/pend variants) to `droid_skills/`
+- **Droid Skills**: Added full skill set (cask/gask/lask/oask + ping/pend variants) to the provider skill assets
 - **Install**: Added `install_droid_skills()` to install Droid skills to `~/.droid/skills/`
 
 </details>
@@ -1110,7 +1139,7 @@ Highlights:
 ### v5.0.6
 - **Zombie Cleanup**: `ccb kill -f` cleans up orphaned tmux sessions globally
 - **Mounted Skill**: Optimized with `pgrep`, extracted to `ccb-mounted` script
-- **Droid Skills**: Full skill set added to `droid_skills/`
+- **Droid Skills**: Full skill set added to the provider skill assets
 
 ### v5.0.5
 - **Droid**: Add delegation tools (`ccb_ask_*` and `cask/gask/lask/oask`) plus `ccb droid setup-delegation` for MCP install

README_zh.md

@@ -7,7 +7,7 @@
   <img src="https://img.shields.io/badge/模型皆可控-CF1322?style=for-the-badge" alt="模型皆可控">
 </p>
 
-[![Version](https://img.shields.io/badge/version-6.2.0-orange.svg)]()
+[![Version](https://img.shields.io/badge/version-6.2.1-orange.svg)]()
 [![Platform](https://img.shields.io/badge/platform-Linux%20%7C%20macOS%20%7C%20Windows-lightgrey.svg)]()
 
 [English](README.md) | **中文**
@@ -74,10 +74,10 @@
 <details>
 <summary><b>最新版本亮点</b></summary>
 
-- **支持 callback ask 链式调用**：正在处理 CCB 任务的 agent 可用 `ccb ask --callback <agent>` 委派必须等待结果的子任务。
-- **Nested ask 路由变为显式规则**：active CCB task 内 plain nested `ask` 会被拒绝；需要结果用 `--callback`，独立无需结果的工作用 `--silence`。
-- **Callback 路由可持久恢复**：CCB 会记录 callback edge，把子任务结果作为 continuation task 交回父 agent，并修复重启窗口里的漏交 continuation。
-- **Ask skill 已补齐工作流说明**：Claude、Codex 和 Droid ask skill 现在记录 callback 委派和提交后停止的规则。
+- **新增继承式 `ccb_config` skill**：Claude 和 Codex 安装现在会继承 CCB 自带的配置设计 skill，用于设计 `.ccb/ccb.config` 和团队记忆。
+- **继承式 skill 统一到 `inherit_skills/`**：CCB 自带 skill 自动继承，`useful_tools/` 保持为用户按需安装的可选工具。
+- **Ask guidance 更短**：CCB 注入简洁英文回复指引，不再给每个 ask body 重复 nested-routing 说明，并识别更多显式完整输出要求。
+- **记忆路由更清晰**：`ccb_config` 角色记忆示例强调 owner-to-next-owner 直接交接，以及并行链路使用独立 root work package。
 
 完整历史见 [新版本记录](#新版本记录)。
 
@@ -103,6 +103,25 @@ tmux 复制粘贴：鼠标左键拖拽即可复制，`Ctrl+Shift+V` 粘贴。
 
 `.ccb/ccb_memory.md` 是项目全局记忆文档。
 
+<details>
+<summary><b>配置设计 Skill</b></summary>
+
+当你希望 agent 帮你设计或更新 CCB 团队，而不是手写配置时，可以使用 `ccb_config`。它会被 Claude 和 Codex 安装继承，重点维护三个用户可编辑文件：
+
+- `.ccb/ccb.config`：团队、provider、pane layout 和 worktree 策略
+- `.ccb/ccb_memory.md`：项目级共享工作流记忆
+- `.ccb/agents/<agent>/memory.md`：每个 agent 的角色记忆
+
+在支持 skill 的 provider 里可以这样调用：
+
+```text
+$ccb_config 为一个 Python library 设计团队：一个 coordinator、两个 worktree 实现 agent、一个 reviewer。
+```
+
+这个 skill 会帮助选择 agent 名称、provider、`inplace` / `git-worktree`、compact layout 语法，以及哪些说明应写入共享记忆或 per-agent memory。它会验证 `.ccb/ccb.config` 是当前配置 authority，并在文件修改完成后提醒你重启 CCB。
+
+</details>
+
 <details>
 <summary><b>布局</b></summary>
 
@@ -311,6 +330,16 @@ ccb reinstall
 历史说明：下面较旧的发布记录里仍可能出现 `askd`、旧 flag 或已移除命令。这些内容仅作为 changelog 历史保留，不代表当前 CLI 入口。
 
 <details open>
+<summary><b>v6.2.1</b> - Inherited CCB Config Skill Release</summary>
+
+- 新增继承式 Claude / Codex `ccb_config` skill，用于设计 `.ccb/ccb.config`、选择 agent 角色/provider/worktree layout，并更新共享和 per-agent memory。
+- 将 CCB 自带继承式 skill 统一放到 `inherit_skills/`，同时保持 `useful_tools/` 为用户按需安装的可选工具，不默认继承。
+- 缩短 ask reply guidance，不再在每个 ask body 注入 nested-routing 说明，注入源文本保持英文，并扩展显式完整输出识别。
+- 简化 project/runtime memory wording，并更新 `ccb_config` memory-routing 示例，强调直接 callback 交接和独立 root work package。
+
+</details>
+
+<details>
 <summary><b>v6.2.0</b> - Callback Ask Chain Release</summary>
 
 - 新增 `ccb ask --callback <agent>`，让 active agent 可委派子任务，并在之后以 continuation task 接收子任务结果。

VERSION

@@ -1 +1 @@
-6.2.0
+6.2.1

ccb

@@ -21,7 +21,7 @@ backend_env = get_backend_env()
 if backend_env and not os.environ.get("CCB_BACKEND_ENV"):
     os.environ["CCB_BACKEND_ENV"] = backend_env
 
-VERSION = "6.2.0"
+VERSION = "6.2.1"
 GIT_COMMIT = "release"
 GIT_DATE = "2026-05-11"
 

docs/ccb-project-shared-memory-plan.md

@@ -138,26 +138,18 @@ communication.
 ````md
 # CCB Project Memory
 
-This project is managed by CCB as a visible multi-agent workspace.
+This project uses CCB for visible multi-agent collaboration.
 
-## Agent Awareness
+## Collaboration
 
 - You are one agent in a CCB-managed project team.
-- Other configured agents may be available in the same project.
-- When work should be handled, reviewed, or cross-checked by another visible
-  project agent, use CCB `ask`.
-- Prefer CCB `ask` over provider-native hidden subagents for project-level
-  collaboration.
-- When delegating, include the goal, relevant files, current assumptions, and
-  expected output.
-- When replying to another agent, be concrete: include findings, changed files,
-  blockers, and verification results.
+- Use CCB `ask` for project-level collaboration with configured agents.
+- Delegate with the goal, scope/files, assumptions, expected output, and verification needs.
+- Reply concisely with findings, changes, verification, blockers, and risks when relevant.
 
 ## Ask Communication
 
-Use CCB `ask` as an asynchronous handoff channel.
-
-Preferred provider form, when slash commands are available:
+Preferred form:
 
 ```text
 /ask <agent> <message>
@@ -171,15 +163,9 @@ $MESSAGE
 EOF
 ```
 
-After submitting, stop. Do not wait for the reply, poll status, or run
-observer commands such as `pend`, `watch`, or `ping` unless the user explicitly
-asks for diagnostics.
-
-When this agent is already handling an active CCB ask task and delegates work
-needed to finish that same task, use `ask --callback` so the child result
-returns as a continuation task. Use `ask --silence` only for independent work
-whose result is not needed by the active task. Plain nested `ask` from an active
-task is rejected by CCB.
+- Submit once, then stop. Do not wait, poll, or run `pend`/`watch`/`ping` unless diagnostics were requested.
+- During an active CCB ask task, use `ask --callback` when a child result is needed to finish; use `ask --silence` only for independent no-result-needed work.
+- Plain nested `ask` from an active task is rejected by CCB.
 ````
 
 ## Memory Bundle Model

docs/current-project-structure.md

@@ -736,8 +736,9 @@ Meaning:
 
 ### Still co-located but not part of the clean core
 
-- `claude_skills/`, `codex_skills/`, `droid_skills/`
-  provider-specific prompt/skill assets
+- `inherit_skills/`
+  provider-specific prompt/skill assets installed by CCB and inherited by
+  managed provider homes
 - `bin/`
   legacy provider wrappers plus current helper scripts
 - `mcp/ccb-delegation/`

claude_skills/ask/RUNTIME.md …erit_skills/claude_skills/ask/RUNTIME.mdclaude_skills/ask/RUNTIME.md renamed to inherit_skills/claude_skills/ask/RUNTIME.md claude_skills/ask/RUNTIME.md renamed to inherit_skills/claude_skills/ask/RUNTIME.md

@@ -8,7 +8,8 @@ metadata:
 Use this only for `/ask <target> <message...>`.
 
 - `TARGET` = first token after `/ask`.
-- `MESSAGE` = exact raw remainder; forward verbatim.
+- `MESSAGE` = exact raw remainder sent as the task body.
+- `ask` may append reply guidance unless the message already contains explicit output requirements.
 - `TARGET=all` broadcasts.
 - Plain `ask` injects concise-reply guidance while still delivering the full reply body.
 - Use `--compact` when the caller wants an actively distilled answer, such as review findings, status, risks, blockers, or next actions.

claude_skills/ask/SKILL.md inherit_skills/claude_skills/ask/SKILL.mdclaude_skills/ask/SKILL.md renamed to inherit_skills/claude_skills/ask/SKILL.md

@@ -7,7 +7,8 @@ metadata:
 
 Use this only for `/ask <target> <message...>`.
 
-- `TARGET` = first token; `MESSAGE` = raw remainder, forwarded verbatim.
+- `TARGET` = first token; `MESSAGE` = raw remainder sent as the task body.
+- `ask` may append reply guidance unless the message already contains explicit output requirements.
 - `TARGET=all` broadcasts.
 - Plain `ask` injects concise-reply guidance while still delivering the full reply body.
 - Use `--compact` when the caller wants an actively distilled answer, such as review findings, status, risks, blockers, or next actions.