docs(zh-CN): sync Chinese docs with latest upstream changes#341
Merged
affaan-m merged 2 commits intoaffaan-m:mainfrom Mar 7, 2026
Merged
docs(zh-CN): sync Chinese docs with latest upstream changes#341affaan-m merged 2 commits intoaffaan-m:mainfrom
affaan-m merged 2 commits intoaffaan-m:mainfrom
Conversation
Contributor
📝 WalkthroughWalkthroughThis pull request adds and updates extensive Chinese-language documentation across agents, commands, skills, hooks, rules, and core docs (v1.8.0 material). Changes are documentation-only; no code, logic, or public API modifications are introduced. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
12 issues found across 53 files
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="docs/zh-CN/skills/continuous-learning-v2/SKILL.md">
<violation number="1" location="docs/zh-CN/skills/continuous-learning-v2/SKILL.md:8">
P2: Heading is broken across two lines. `的架构` was accidentally split from the `#` heading onto a separate line (as a plain paragraph), so the rendered title reads "基于本能" and a stray "的架构" appears below it. These should be on one line.</violation>
</file>
<file name="docs/zh-CN/commands/instinct-import.md">
<violation number="1" location="docs/zh-CN/commands/instinct-import.md:112">
P2: The output example still shows the global instincts path (`~/.claude/homunculus/instincts/inherited/`), but the new `--scope` flag defaults to `project`. The example output should show the project-scoped path (e.g., `~/.claude/homunculus/projects/<project-id>/instincts/inherited/`) to be consistent with the default behavior documented elsewhere in this file.</violation>
</file>
<file name="docs/zh-CN/skills/autonomous-loops/SKILL.md">
<violation number="1" location="docs/zh-CN/skills/autonomous-loops/SKILL.md:241">
P2: This `curl | bash` install command references an external, unvetted repository (`AnandChowdhary/continuous-claude`). This is both a supply-chain risk (the remote script could change at any time) and follows the `curl | bash` anti-pattern. Consider either porting the installation into this repo, linking to the repository without the pipe-to-bash pattern, or adding a note advising users to inspect the script before running it.
(Based on your team's feedback about avoiding linking to unvetted external repositories.) [FEEDBACK_USED]</violation>
</file>
<file name="docs/zh-CN/skills/ralphinho-rfc-pipeline/SKILL.md">
<violation number="1" location="docs/zh-CN/skills/ralphinho-rfc-pipeline/SKILL.md:9">
P2: This links to an external repository (`https://github.com/humanplane`) that may not be vetted by the organization. Consider removing the hyperlink and keeping it as plain text, or verifying the repo is organization-controlled. Note: the English source file has the same link and should be addressed there as well.
(Based on your team's feedback about avoiding linking to unvetted external repositories.) [FEEDBACK_USED]</violation>
</file>
<file name="docs/zh-CN/commands/instinct-status.md">
<violation number="1" location="docs/zh-CN/commands/instinct-status.md:3">
P2: Mistranslation: "并充满信心" means "and be full of confidence" (emotional state), but the English source "with confidence" means "with confidence levels/scores" (a technical metric). This should be "及其置信度" or "并显示置信度" to match the upstream meaning.</violation>
</file>
<file name="docs/zh-CN/agents/harness-optimizer.md">
<violation number="1" location="docs/zh-CN/agents/harness-optimizer.md:9">
P2: "线束" is a mistranslation of "harness" in this software context — it literally means "wire harness" (electrical engineering). The rest of the zh-CN docs translate "harness" as "框架" (framework), e.g. `agent-harness-construction` → "智能体框架构建". This is also inconsistent with this file's own frontmatter description which uses "代理工具配置". Consider using "框架" for consistency.</violation>
<violation number="2" location="docs/zh-CN/agents/harness-optimizer.md:13">
P2: Same mistranslation of "harness" as "线束" — should be "框架" to match the rest of the zh-CN documentation.</violation>
</file>
<file name="docs/zh-CN/CODE_OF_CONDUCT.md">
<violation number="1" location="docs/zh-CN/CODE_OF_CONDUCT.md:39">
P2: Missing contact email address. The sentence says "邮箱为。" ("email is.") but no email address is provided. Either add the actual contact email or restructure the sentence (e.g., match the English phrasing that omits "email").</violation>
<violation number="2" location="docs/zh-CN/CODE_OF_CONDUCT.md:74">
P2: Escaped brackets `\[...]\[...]` break the markdown reference link. This will render as literal text instead of a clickable link to contributor-covenant.org. Remove the backslash escapes to restore proper link rendering.</violation>
<violation number="3" location="docs/zh-CN/CODE_OF_CONDUCT.md:76">
P2: Escaped underscores in the URL (`code\_of\_conduct.html`) will break the link. This is likely an artifact from a markdown conversion tool. Remove the backslash escapes.</violation>
</file>
<file name="docs/zh-CN/agents/loop-operator.md">
<violation number="1" location="docs/zh-CN/agents/loop-operator.md:17">
P2: Translation error: both "pattern" and "mode" are rendered as the same word "模式", making the sentence read "从明确的**模式**和**模式**开始循环" ("from explicit pattern and pattern"). This loses the distinction from the original English. Consider differentiating them, e.g. using "匹配模式" for pattern and "运行模式" for mode.</violation>
</file>
<file name="docs/zh-CN/skills/configure-ecc/SKILL.md">
<violation number="1" location="docs/zh-CN/skills/configure-ecc/SKILL.md:100">
P2: Mistranslation: `个人技能` means "personal skills" (relating to a person), but the section is about confirming specific/individual skill selections. The original `单项技能` (individual skills) was correct. Consider reverting to `确认单项技能` or using `确认各项技能`.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
Comment on lines
+8
to
+10
| # 持续学习 v2.1 - 基于本能 | ||
|
|
||
| 的架构 |
There was a problem hiding this comment.
P2: Heading is broken across two lines. 的架构 was accidentally split from the # heading onto a separate line (as a plain paragraph), so the rendered title reads "基于本能" and a stray "的架构" appears below it. These should be on one line.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/skills/continuous-learning-v2/SKILL.md, line 8:
<comment>Heading is broken across two lines. `的架构` was accidentally split from the `#` heading onto a separate line (as a plain paragraph), so the rendered title reads "基于本能" and a stray "的架构" appears below it. These should be on one line.</comment>
<file context>
@@ -1,33 +1,48 @@
---
-# 持续学习 v2 - 基于本能的架构
+# 持续学习 v2.1 - 基于本能
+
+的架构
</file context>
Suggested change
| # 持续学习 v2.1 - 基于本能 | |
| 的架构 | |
| # 持续学习 v2.1 - 基于本能的架构 |
| Added: 8 instincts | ||
| Updated: 1 instinct | ||
| Skipped: 3 instincts (2 duplicates, 1 conflict) | ||
| Skipped: 3 instincts (equal/higher confidence already exists) |
There was a problem hiding this comment.
P2: The output example still shows the global instincts path (~/.claude/homunculus/instincts/inherited/), but the new --scope flag defaults to project. The example output should show the project-scoped path (e.g., ~/.claude/homunculus/projects/<project-id>/instincts/inherited/) to be consistent with the default behavior documented elsewhere in this file.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/commands/instinct-import.md, line 112:
<comment>The output example still shows the global instincts path (`~/.claude/homunculus/instincts/inherited/`), but the new `--scope` flag defaults to `project`. The example output should show the project-scoped path (e.g., `~/.claude/homunculus/projects/<project-id>/instincts/inherited/`) to be consistent with the default behavior documented elsewhere in this file.</comment>
<file context>
@@ -142,7 +109,7 @@ original_source: "session-observation" # or "repo-analysis"
Added: 8 instincts
Updated: 1 instinct
-Skipped: 3 instincts (2 duplicates, 1 conflict)
+Skipped: 3 instincts (equal/higher confidence already exists)
New instincts saved to: ~/.claude/homunculus/instincts/inherited/
</file context>
| ### 安装 | ||
|
|
||
| ```bash | ||
| curl -fsSL https://raw.githubusercontent.com/AnandChowdhary/continuous-claude/HEAD/install.sh | bash |
There was a problem hiding this comment.
P2: This curl | bash install command references an external, unvetted repository (AnandChowdhary/continuous-claude). This is both a supply-chain risk (the remote script could change at any time) and follows the curl | bash anti-pattern. Consider either porting the installation into this repo, linking to the repository without the pipe-to-bash pattern, or adding a note advising users to inspect the script before running it.
(Based on your team's feedback about avoiding linking to unvetted external repositories.)
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/skills/autonomous-loops/SKILL.md, line 241:
<comment>This `curl | bash` install command references an external, unvetted repository (`AnandChowdhary/continuous-claude`). This is both a supply-chain risk (the remote script could change at any time) and follows the `curl | bash` anti-pattern. Consider either porting the installation into this repo, linking to the repository without the pipe-to-bash pattern, or adding a note advising users to inspect the script before running it.
(Based on your team's feedback about avoiding linking to unvetted external repositories.) </comment>
<file context>
@@ -0,0 +1,621 @@
+### 安装
+
+```bash
+curl -fsSL https://raw.githubusercontent.com/AnandChowdhary/continuous-claude/HEAD/install.sh | bash
+```
+
</file context>
|
|
||
| # Ralphinho RFC 管道 | ||
|
|
||
| 灵感来源于 [humanplane](https://github.com/humanplane) 风格的 RFC 分解模式和多单元编排工作流。 |
There was a problem hiding this comment.
P2: This links to an external repository (https://github.com/humanplane) that may not be vetted by the organization. Consider removing the hyperlink and keeping it as plain text, or verifying the repo is organization-controlled. Note: the English source file has the same link and should be addressed there as well.
(Based on your team's feedback about avoiding linking to unvetted external repositories.)
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/skills/ralphinho-rfc-pipeline/SKILL.md, line 9:
<comment>This links to an external repository (`https://github.com/humanplane`) that may not be vetted by the organization. Consider removing the hyperlink and keeping it as plain text, or verifying the repo is organization-controlled. Note: the English source file has the same link and should be addressed there as well.
(Based on your team's feedback about avoiding linking to unvetted external repositories.) </comment>
<file context>
@@ -0,0 +1,69 @@
+
+# Ralphinho RFC 管道
+
+灵感来源于 [humanplane](https://github.com/humanplane) 风格的 RFC 分解模式和多单元编排工作流。
+
+当一个功能对于单次代理处理来说过于庞大，必须拆分为独立可验证的工作单元时，请使用此技能。
</file context>
Suggested change
| 灵感来源于 [humanplane](https://github.com/humanplane) 风格的 RFC 分解模式和多单元编排工作流。 | |
| 灵感来源于 humanplane 风格的 RFC 分解模式和多单元编排工作流。 |
| --- | ||
| name: instinct-status | ||
| description: 显示所有已学习的本能及其置信水平 | ||
| description: 展示已学习的本能（项目+全局）并充满信心 |
There was a problem hiding this comment.
P2: Mistranslation: "并充满信心" means "and be full of confidence" (emotional state), but the English source "with confidence" means "with confidence levels/scores" (a technical metric). This should be "及其置信度" or "并显示置信度" to match the upstream meaning.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/commands/instinct-status.md, line 3:
<comment>Mistranslation: "并充满信心" means "and be full of confidence" (emotional state), but the English source "with confidence" means "with confidence levels/scores" (a technical metric). This should be "及其置信度" or "并显示置信度" to match the upstream meaning.</comment>
<file context>
@@ -1,12 +1,12 @@
---
name: instinct-status
-description: 显示所有已学习的本能及其置信水平
+description: 展示已学习的本能（项目+全局）并充满信心
command: true
---
</file context>
Suggested change
| description: 展示已学习的本能（项目+全局）并充满信心 | |
| description: 展示已学习的本能（项目+全局）及其置信度 |
|
|
||
| ## 归属 | ||
|
|
||
| 本《行为准则》改编自 \[Contributor Covenant]\[homepage]， |
There was a problem hiding this comment.
P2: Escaped brackets \[...]\[...] break the markdown reference link. This will render as literal text instead of a clickable link to contributor-covenant.org. Remove the backslash escapes to restore proper link rendering.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/CODE_OF_CONDUCT.md, line 74:
<comment>Escaped brackets `\[...]\[...]` break the markdown reference link. This will render as literal text instead of a clickable link to contributor-covenant.org. Remove the backslash escapes to restore proper link rendering.</comment>
<file context>
@@ -0,0 +1,84 @@
+
+## 归属
+
+本《行为准则》改编自 \[Contributor Covenant]\[homepage]，
+版本 2.0，可在
+https://www.contributor-covenant.org/version/2/0/code\_of\_conduct.html 获取。
</file context>
Suggested change
| 本《行为准则》改编自 \[Contributor Covenant]\[homepage]， | |
| 本《行为准则》改编自 [Contributor Covenant][homepage]， |
|
|
||
| ## 执行 | ||
|
|
||
| 辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告，邮箱为。 |
There was a problem hiding this comment.
P2: Missing contact email address. The sentence says "邮箱为。" ("email is.") but no email address is provided. Either add the actual contact email or restructure the sentence (e.g., match the English phrasing that omits "email").
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/CODE_OF_CONDUCT.md, line 39:
<comment>Missing contact email address. The sentence says "邮箱为。" ("email is.") but no email address is provided. Either add the actual contact email or restructure the sentence (e.g., match the English phrasing that omits "email").</comment>
<file context>
@@ -0,0 +1,84 @@
+
+## 执行
+
+辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告，邮箱为。
+所有投诉都将得到及时和公正的审查和调查。
+
</file context>
Suggested change
| 辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告，邮箱为。 | |
| 辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告。 |
|
|
||
| 本《行为准则》改编自 \[Contributor Covenant]\[homepage]， | ||
| 版本 2.0，可在 | ||
| https://www.contributor-covenant.org/version/2/0/code\_of\_conduct.html 获取。 |
There was a problem hiding this comment.
P2: Escaped underscores in the URL (code\_of\_conduct.html) will break the link. This is likely an artifact from a markdown conversion tool. Remove the backslash escapes.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/CODE_OF_CONDUCT.md, line 76:
<comment>Escaped underscores in the URL (`code\_of\_conduct.html`) will break the link. This is likely an artifact from a markdown conversion tool. Remove the backslash escapes.</comment>
<file context>
@@ -0,0 +1,84 @@
+
+本《行为准则》改编自 \[Contributor Covenant]\[homepage]，
+版本 2.0，可在
+https://www.contributor-covenant.org/version/2/0/code\_of\_conduct.html 获取。
+
+社区影响指南的灵感来源于 [Mozilla 的行为准则执行阶梯](https://github.com/mozilla/diversity)。
</file context>
Suggested change
| https://www.contributor-covenant.org/version/2/0/code\_of\_conduct.html 获取。 | |
| https://www.contributor-covenant.org/version/2/0/code_of_conduct.html 获取。 |
|
|
||
| ## 工作流程 | ||
|
|
||
| 1. 从明确的模式和模式开始循环。 |
There was a problem hiding this comment.
P2: Translation error: both "pattern" and "mode" are rendered as the same word "模式", making the sentence read "从明确的模式和模式开始循环" ("from explicit pattern and pattern"). This loses the distinction from the original English. Consider differentiating them, e.g. using "匹配模式" for pattern and "运行模式" for mode.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/agents/loop-operator.md, line 17:
<comment>Translation error: both "pattern" and "mode" are rendered as the same word "模式", making the sentence read "从明确的**模式**和**模式**开始循环" ("from explicit pattern and pattern"). This loses the distinction from the original English. Consider differentiating them, e.g. using "匹配模式" for pattern and "运行模式" for mode.</comment>
<file context>
@@ -0,0 +1,37 @@
+
+## 工作流程
+
+1. 从明确的模式和模式开始循环。
+2. 跟踪进度检查点。
+3. 检测停滞和重试风暴。
</file context>
| ``` | ||
|
|
||
| ### 2b：确认单项技能 | ||
| ### 2c: 确认个人技能 |
There was a problem hiding this comment.
P2: Mistranslation: 个人技能 means "personal skills" (relating to a person), but the section is about confirming specific/individual skill selections. The original 单项技能 (individual skills) was correct. Consider reverting to 确认单项技能 or using 确认各项技能.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At docs/zh-CN/skills/configure-ecc/SKILL.md, line 100:
<comment>Mistranslation: `个人技能` means "personal skills" (relating to a person), but the section is about confirming specific/individual skill selections. The original `单项技能` (individual skills) was correct. Consider reverting to `确认单项技能` or using `确认各项技能`.</comment>
<file context>
@@ -80,7 +97,7 @@ Options:
-### 2b：确认单项技能
+### 2c: 确认个人技能
对于每个选定的类别，打印下面的完整技能列表，并要求用户确认或取消选择特定的技能。如果列表超过 4 项，将列表打印为文本，并使用 AskUserQuestion，提供一个 "安装所有列出项" 的选项，以及一个 "其他" 选项供用户粘贴特定名称。
</file context>
</details>
<a href="https://www.cubic.dev/action/fix/violation/06c84c50-05ae-4919-b016-c0f9ae489062" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://cubic.dev/buttons/fix-with-cubic-dark.svg">
<source media="(prefers-color-scheme: light)" srcset="https://cubic.dev/buttons/fix-with-cubic-light.svg">
<img alt="Fix with Cubic" src="https://cubic.dev/buttons/fix-with-cubic-dark.svg">
</picture>
</a>
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 15
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (4)
docs/zh-CN/skills/configure-ecc/SKILL.md (2)
120-120:⚠️ Potential issue | 🟡 Minor
fixturesis mistranslated here.
固件means firmware, not pytest fixtures. This changes the meaning of the skill description and should use the usual testing term instead.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/skills/configure-ecc/SKILL.md` at line 120, The skill description for `python-testing` mistranslates "fixtures" as "固件" (firmware); edit the `python-testing` entry to replace "固件" with the correct testing term such as "测试夹具" (or "fixtures/测试夹具") so the line reads e.g. "使用 pytest、TDD、测试夹具、模拟、参数化进行 Python 测试", keeping the rest of the phrase unchanged.
89-97:⚠️ Potential issue | 🟠 MajorStep 2b drops categories that 2c still asks users to confirm.
Line 89 says there are 27 skills across 4 categories, but the table below enumerates 34 skills across 5 groups, and
Business & ContentplusStandaloneare missing from the selection options. That makesarticle-writing,content-engine,market-research,investor-*, andproject-guidelines-exampleunreachable through the documented flow.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/skills/configure-ecc/SKILL.md` around lines 89 - 97, Update the "Which skill categories do you want to install?" question and its Options so they match the table below: correct the total count (27 → 34) and add the missing categories "Business & Content" and "Standalone" to the Options list, ensuring entries map to the unreachable skills (article-writing, content-engine, market-research, investor-*, project-guidelines-example); modify the Options block under the Question (the exact block starting with Question: "Which skill categories do you want to install?") to include those two categories and adjust descriptions so all five groups are represented and selectable.docs/zh-CN/commands/instinct-import.md (1)
103-115:⚠️ Potential issue | 🟡 MinorMake the completion path example scope-aware.
The command now defaults to
--scope project, but this sample output still always prints the global inherited path. That will mislead readers about where imports land unless they explicitly pass--scope global.Suggested doc fix
-New instincts saved to: ~/.claude/homunculus/instincts/inherited/ +New instincts saved to: ~/.claude/homunculus/projects/<project-id>/instincts/inherited/ + +# With --scope global: +# New instincts saved to: ~/.claude/homunculus/instincts/inherited/🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/commands/instinct-import.md` around lines 103 - 115, Update the example output under "## 输出" so the completion path is scope-aware: mention the default "--scope project" and either show the project-scoped path (e.g. the location where instincts are saved for project scope) or present two example outputs (one for "--scope project" and one for "--scope global") instead of always printing "~/.claude/homunculus/instincts/inherited/"; modify the text that follows the status lines ("Added/Updated/Skipped") to conditionally reflect the chosen scope or explicitly state which scope the shown path corresponds to.docs/zh-CN/skills/continuous-learning-v2/SKILL.md (1)
147-165:⚠️ Potential issue | 🟠 MajorPass the hook phase explicitly here.
observe.shdefaults topostwhen no argument is provided, so both of these snippets currently documentPreToolUseas a post-event. That will misclassify observations and undermine the “PreToolUse/PostToolUse (100% reliable)” behavior described above.Proposed fix
"PreToolUse": [{ "matcher": "*", "hooks": [{ "type": "command", - "command": "${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning-v2/hooks/observe.sh" + "command": "${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning-v2/hooks/observe.sh pre" }] }], "PostToolUse": [{ "matcher": "*", "hooks": [{ "type": "command", - "command": "${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning-v2/hooks/observe.sh" + "command": "${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning-v2/hooks/observe.sh post" }] }]"PreToolUse": [{ "matcher": "*", "hooks": [{ "type": "command", - "command": "~/.claude/skills/continuous-learning-v2/hooks/observe.sh" + "command": "~/.claude/skills/continuous-learning-v2/hooks/observe.sh pre" }] }], "PostToolUse": [{ "matcher": "*", "hooks": [{ "type": "command", - "command": "~/.claude/skills/continuous-learning-v2/hooks/observe.sh" + "command": "~/.claude/skills/continuous-learning-v2/hooks/observe.sh post" }] }]Also applies to: 170-188
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/skills/continuous-learning-v2/SKILL.md` around lines 147 - 165, The hooks currently call observe.sh with no arguments (command: "${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning-v2/hooks/observe.sh"), but observe.sh defaults to post, so update the PreToolUse and PostToolUse hook command entries to pass the explicit phase argument: call observe.sh "pre" for PreToolUse and "post" for PostToolUse (leave matcher and hook structure intact) so observations are correctly classified by the PreToolUse and PostToolUse hooks.
🧹 Nitpick comments (1)
docs/zh-CN/skills/agent-harness-construction/SKILL.md (1)
7-77: Align this skill with the repo’s standard section layout.The content is solid, but this doc skips the usual
## 何时使用/## 工作原理/## 示例structure the repo uses for skills. Adding those headings would make it easier to scan and keep consistent with other skill docs.Based on learnings: Applies to skills/**/*.md : Skills should be formatted as Markdown with clear sections for When to Use, How It Works, and Examples
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/skills/agent-harness-construction/SKILL.md` around lines 7 - 77, The doc needs the repo's standard skill layout: add the three sections "## 何时使用", "## 工作原理", and "## 示例" to SKILL.md, moving existing high-level guidance into "## 工作原理", summarizing practical scenarios under "## 何时使用" (e.g., when improving planning, tool-calls, recovery, context budgeting), and creating 1–2 concise concrete use examples under "## 示例" that reference the Core Model, Action Space, Observation, and Error Recovery sections; keep original content but reorganize into these headings for consistency.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@docs/zh-CN/AGENTS.md`:
- Around line 3-4: Update stale headline counts: the intro line that currently
reads "13 个智能体、50+ 项技能、33 条命令" should be changed to "16 个智能体、65 项技能和 40 条命令";
also update the matching counts inside the "project-structure" block (the
translated project structure section) and the other occurrences referenced
around lines 125-127 so the doc's translated counts match docs/zh-CN/README.md;
locate occurrences by searching for the current numeric string "13 个智能体" / "50+
项技能" / "33 条命令" and replace them with the new values.
In `@docs/zh-CN/agents/code-reviewer.md`:
- Around line 230-233: Replace the awkward phrase "不必要的增加模型成本的复杂性" in the list
item with a more natural wording; update the item text to either
"不必要地增加模型成本的复杂性" or "无谓抬高模型成本的复杂性" (choose one) so the sentence reads smoothly
and grammatically correct.
In `@docs/zh-CN/CLAUDE.md`:
- Line 7: The sentence that calls the repo "这是一个 **Claude Code 插件**" is
misleading; update the opening line in CLAUDE.md to describe the repository as a
collection of components (agents, skills, hooks, commands, rules, and MCP
configs) rather than a single plugin—replace the phrase "Claude Code 插件" with
wording such as "一组 Claude Code 组件/资源" or "包含代理、技能、钩子、命令、规则和 MCP 配置的集合" and
ensure the rest of the sentence still conveys the production-ready workflows for
using Claude Code.
In `@docs/zh-CN/CODE_OF_CONDUCT.md`:
- Around line 39-40: The enforcement contact is missing in the sentence that
currently reads "辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告，邮箱为。"; update that line in
CODE_OF_CONDUCT.md to include a valid enforcement email address or a placeholder
(e.g., [email protected] or [email protected]) so contributors know where to
send reports, and ensure the updated sentence keeps the same wording and
punctuation around the added address.
In `@docs/zh-CN/commands/instinct-export.md`:
- Around line 22-23: The docs show an /instinct-export example but don’t call
out that export now defaults to --scope all while /instinct-import still
defaults to project, so add an explicit note and matching import example that
sets the same --scope (e.g., use --scope project in both export and import or
--scope all in both) to preserve intent; update the example block with
/instinct-import and add a brief warning sentence about the default-scope
mismatch and mirror this change in the other example ranges referenced (the
example pairs around the blocks currently at lines 27-33 and 64-67) so readers
see a round-trip that preserves scope.
In `@docs/zh-CN/commands/loop-start.md`:
- Around line 24-26: Update the wording to clarify hook safety: state that
ECC_HOOK_PROFILE falls back to "standard" on invalid values (do not say users
"disable globally") and instruct readers to both verify the selected
ECC_HOOK_PROFILE and confirm that required safety hooks are not listed in
ECC_DISABLED_HOOKS so the safety checks remain active (see adapter behavior
referenced in .cursor/hooks/adapter.js:63-79 and hooks/README.md:73-87).
In `@docs/zh-CN/commands/model-route.md`:
- Around line 15-20: Replace the awkward heading text "必需输出" with the natural
phrasing "必须输出" in the model-route documentation; locate the section whose
heading currently reads "必需输出" (the one listing 推荐的模型, 置信度, 该模型适合的原因,
如果首次尝试失败，备用的回退模型) and update that heading string to "必须输出".
In `@docs/zh-CN/commands/multi-workflow.md`:
- Around line 16-18: Replace the inconsistent Chinese term "质量关卡" in the bullet
that currently reads "结构化的 6 阶段工作流程，带有质量关卡" with the established term family
"质量门控" (optionally add the English parenthetical "quality-gate") so the line
becomes e.g. "结构化的 6 阶段工作流程，带有质量门控（quality-gate）", ensuring consistency with
other docs that use "质量门控"/"quality-gate".
In `@docs/zh-CN/hooks/README.md`:
- Around line 143-153: The example JSON shown for the asynchronous hook is
missing the required outer hooks shape; update the snippet so it is either
wrapped in the standard hooks.json structure (include a "matcher" object and a
"hooks" array with the inner hook containing "type": "command", "command": "node
my-slow-hook.js", "async": true, "timeout": 30) or clearly label it as just the
inner hook item; reference the "matcher" and "hooks" wrapper and the inner keys
("type", "command", "async", "timeout") when making the change.
In `@docs/zh-CN/README.md`:
- Around line 8-10: The hardcoded "150 installs" in the README's Shields badge
will drift; locate the Markdown line containing the "GitHub App Install" badge
(the shields URL and link) and replace it with a non-numeric or count-free
alternative—either remove "150 installs" from the badge label (e.g., "GitHub
App") or swap the badge for a generic Marketplace link/button so the README no
longer contains a hardcoded install count.
- Around line 615-623: The manual install commands in README.md currently copy
from everything-claude-code/.agents/skills/* which is incorrect for Claude Code
(the repo documents Claude Code skills under skills/); update the copy commands
to use everything-claude-code/skills/ (and keep the separate copy for
everything-claude-code/skills/search-first) so users copy the correct Claude
Code skill format instead of the Codex .agents/skills format; update or add a
brief note near the cp commands clarifying that .agents/skills/ is for Codex
only if you want to keep that alternative mentioned.
In `@docs/zh-CN/skills/configure-ecc/SKILL.md`:
- Around line 72-82: The Chinese doc replaced canonical skill paths/IDs with
generic labels (e.g., `.agents/skills/*`, `research-first`, `evals`) causing the
"core" preset to be ambiguous; update the block that defines the core preset and
the AskUserQuestion options to use the concrete, canonical skill directories and
IDs from the English source (use `$ECC_ROOT/skills/<skill-name>` style names) so
the install flow maps to stable installable skills (e.g., list the actual skill
folder names under `$ECC_ROOT/skills/` instead of labels like `research-first`),
and ensure the AskUserQuestion options and Default still reference the canonical
"Core only (recommended)" set of skill IDs.
In `@docs/zh-CN/skills/plankton-code-quality/SKILL.md`:
- Around line 201-243: This "ECC v1.8 新增内容" block mixes ECC-specific features
into the Plankton skill doc; either remove it or clearly mark it as ECC
integration. Fix by moving the entire section into the existing "Pairing with
ECC" area (refer to the "Pairing with ECC" section), or rename the header to
"Integration with ECC v1.8" and annotate each subsection (e.g., "Language gate
table", the env vars
ECC_HOOK_PROFILE/ECC_QUALITY_GATE_FIX/ECC_QUALITY_GATE_STRICT, "配置篡改防护", "CI
集成模式", "健康指标") to explicitly state these are ECC features and reconcile the
listed config files with Plankton's protection list.
In `@docs/zh-CN/skills/postgres-patterns/SKILL.md`:
- Line 154: Restore the upstream source link in the attribution line by editing
the string '*基于 Supabase 代理技能（致谢：Supabase 团队）（MIT 许可证）*' to include the original
upstream reference URL (e.g. the Supabase repo or source file link) while
keeping the Chinese localization intact; ensure the final line clearly contains
both the localized attribution and the original source link so provenance and
license verification remain visible.
In `@docs/zh-CN/the-longform-guide.md`:
- Around line 139-140: Replace the awkward Chinese sentence in the image caption
describing the mgrep benchmark so it reads naturally and unambiguously: update
the sentence mentioning "mgrep + Claude Code" and the token usage comparison to
state that it uses about half the tokens of the grep-based workflow (e.g., "使用的
token 约为基于 grep 的工作流的一半" or "节省了约一半的 token") and keep the source attribution
"来源：@mixedbread-ai 的 mgrep"; ensure the caption remains a single fluent sentence
in Chinese.
---
Outside diff comments:
In `@docs/zh-CN/commands/instinct-import.md`:
- Around line 103-115: Update the example output under "## 输出" so the completion
path is scope-aware: mention the default "--scope project" and either show the
project-scoped path (e.g. the location where instincts are saved for project
scope) or present two example outputs (one for "--scope project" and one for
"--scope global") instead of always printing
"~/.claude/homunculus/instincts/inherited/"; modify the text that follows the
status lines ("Added/Updated/Skipped") to conditionally reflect the chosen scope
or explicitly state which scope the shown path corresponds to.
In `@docs/zh-CN/skills/configure-ecc/SKILL.md`:
- Line 120: The skill description for `python-testing` mistranslates "fixtures"
as "固件" (firmware); edit the `python-testing` entry to replace "固件" with the
correct testing term such as "测试夹具" (or "fixtures/测试夹具") so the line reads e.g.
"使用 pytest、TDD、测试夹具、模拟、参数化进行 Python 测试", keeping the rest of the phrase
unchanged.
- Around line 89-97: Update the "Which skill categories do you want to install?"
question and its Options so they match the table below: correct the total count
(27 → 34) and add the missing categories "Business & Content" and "Standalone"
to the Options list, ensuring entries map to the unreachable skills
(article-writing, content-engine, market-research, investor-*,
project-guidelines-example); modify the Options block under the Question (the
exact block starting with Question: "Which skill categories do you want to
install?") to include those two categories and adjust descriptions so all five
groups are represented and selectable.
In `@docs/zh-CN/skills/continuous-learning-v2/SKILL.md`:
- Around line 147-165: The hooks currently call observe.sh with no arguments
(command:
"${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning-v2/hooks/observe.sh"), but
observe.sh defaults to post, so update the PreToolUse and PostToolUse hook
command entries to pass the explicit phase argument: call observe.sh "pre" for
PreToolUse and "post" for PostToolUse (leave matcher and hook structure intact)
so observations are correctly classified by the PreToolUse and PostToolUse
hooks.
---
Nitpick comments:
In `@docs/zh-CN/skills/agent-harness-construction/SKILL.md`:
- Around line 7-77: The doc needs the repo's standard skill layout: add the
three sections "## 何时使用", "## 工作原理", and "## 示例" to SKILL.md, moving existing
high-level guidance into "## 工作原理", summarizing practical scenarios under "##
何时使用" (e.g., when improving planning, tool-calls, recovery, context budgeting),
and creating 1–2 concise concrete use examples under "## 示例" that reference the
Core Model, Action Space, Observation, and Error Recovery sections; keep
original content but reorganize into these headings for consistency.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 87d7e4f4-3b02-4b84-8554-5cc224d75305
📒 Files selected for processing (53)
docs/zh-CN/AGENTS.mddocs/zh-CN/CHANGELOG.mddocs/zh-CN/CLAUDE.mddocs/zh-CN/CODE_OF_CONDUCT.mddocs/zh-CN/CONTRIBUTING.mddocs/zh-CN/README.mddocs/zh-CN/SPONSORING.mddocs/zh-CN/SPONSORS.mddocs/zh-CN/agents/chief-of-staff.mddocs/zh-CN/agents/code-reviewer.mddocs/zh-CN/agents/database-reviewer.mddocs/zh-CN/agents/harness-optimizer.mddocs/zh-CN/agents/loop-operator.mddocs/zh-CN/agents/tdd-guide.mddocs/zh-CN/commands/claw.mddocs/zh-CN/commands/evolve.mddocs/zh-CN/commands/harness-audit.mddocs/zh-CN/commands/instinct-export.mddocs/zh-CN/commands/instinct-import.mddocs/zh-CN/commands/instinct-status.mddocs/zh-CN/commands/loop-start.mddocs/zh-CN/commands/loop-status.mddocs/zh-CN/commands/model-route.mddocs/zh-CN/commands/multi-backend.mddocs/zh-CN/commands/multi-execute.mddocs/zh-CN/commands/multi-frontend.mddocs/zh-CN/commands/multi-plan.mddocs/zh-CN/commands/multi-workflow.mddocs/zh-CN/commands/projects.mddocs/zh-CN/commands/promote.mddocs/zh-CN/commands/quality-gate.mddocs/zh-CN/hooks/README.mddocs/zh-CN/plugins/README.mddocs/zh-CN/rules/common/development-workflow.mddocs/zh-CN/skills/agent-harness-construction/SKILL.mddocs/zh-CN/skills/agentic-engineering/SKILL.mddocs/zh-CN/skills/ai-first-engineering/SKILL.mddocs/zh-CN/skills/autonomous-loops/SKILL.mddocs/zh-CN/skills/configure-ecc/SKILL.mddocs/zh-CN/skills/continuous-agent-loop/SKILL.mddocs/zh-CN/skills/continuous-learning-v2/SKILL.mddocs/zh-CN/skills/continuous-learning-v2/agents/observer.mddocs/zh-CN/skills/continuous-learning/SKILL.mddocs/zh-CN/skills/cpp-testing/SKILL.mddocs/zh-CN/skills/enterprise-agent-ops/SKILL.mddocs/zh-CN/skills/eval-harness/SKILL.mddocs/zh-CN/skills/frontend-slides/SKILL.mddocs/zh-CN/skills/nanoclaw-repl/SKILL.mddocs/zh-CN/skills/plankton-code-quality/SKILL.mddocs/zh-CN/skills/postgres-patterns/SKILL.mddocs/zh-CN/skills/ralphinho-rfc-pipeline/SKILL.mddocs/zh-CN/skills/search-first/SKILL.mddocs/zh-CN/the-longform-guide.md
Comment on lines
+3
to
+4
| 这是一个**生产就绪的 AI 编码插件**，提供 13 个专业智能体、50+ 项技能、33 条命令以及用于软件开发的自动化钩子工作流。 | ||
|
|
Contributor
There was a problem hiding this comment.
The headline counts are stale and now contradict the zh-CN README.
This file still says 13/50+/33, while docs/zh-CN/README.md now advertises 16 个智能体、65 项技能和 40 条命令. Please update both the intro and the project-structure block so the translated docs stay internally consistent.
Also applies to: 125-127
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/AGENTS.md` around lines 3 - 4, Update stale headline counts: the
intro line that currently reads "13 个智能体、50+ 项技能、33 条命令" should be changed to
"16 个智能体、65 项技能和 40 条命令"; also update the matching counts inside the
"project-structure" block (the translated project structure section) and the
other occurrences referenced around lines 125-127 so the doc's translated counts
match docs/zh-CN/README.md; locate occurrences by searching for the current
numeric string "13 个智能体" / "50+ 项技能" / "33 条命令" and replace them with the new
values.
Comment on lines
+230
to
+233
| 1. 行为回归和边缘情况处理 | ||
| 2. 安全假设和信任边界 | ||
| 3. 隐藏的耦合或意外的架构漂移 | ||
| 4. 不必要的增加模型成本的复杂性 |
Contributor
There was a problem hiding this comment.
Tighten the wording of item 4.
不必要的增加模型成本的复杂性 reads awkwardly. A more natural phrasing would be 不必要地增加模型成本的复杂性 or 无谓抬高模型成本的复杂性.
✏️ Suggested wording
-4. 不必要的增加模型成本的复杂性
+4. 不必要地增加模型成本的复杂性📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| 1. 行为回归和边缘情况处理 | |
| 2. 安全假设和信任边界 | |
| 3. 隐藏的耦合或意外的架构漂移 | |
| 4. 不必要的增加模型成本的复杂性 | |
| 1. 行为回归和边缘情况处理 | |
| 2. 安全假设和信任边界 | |
| 3. 隐藏的耦合或意外的架构漂移 | |
| 4. 不必要地增加模型成本的复杂性 |
🧰 Tools
🪛 LanguageTool
[uncategorized] ~233-~233: 动词的修饰一般为‘形容词（副词）+地+动词’。您的意思是否是：不必要"地"增加
Context: ...况处理 2. 安全假设和信任边界 3. 隐藏的耦合或意外的架构漂移 4. 不必要的增加模型成本的复杂性 成本意识检查： * 标记那些在没有明确理由需求的情况下...
(wb4)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/agents/code-reviewer.md` around lines 230 - 233, Replace the
awkward phrase "不必要的增加模型成本的复杂性" in the list item with a more natural wording;
update the item text to either "不必要地增加模型成本的复杂性" or "无谓抬高模型成本的复杂性" (choose one)
so the sentence reads smoothly and grammatically correct.
|
|
||
| ## 项目概述 | ||
|
|
||
| 这是一个 **Claude Code 插件** - 一个包含生产就绪的代理、技能、钩子、命令、规则和 MCP 配置的集合。该项目提供了使用 Claude Code 进行软件开发的经验证的工作流。 |
Contributor
There was a problem hiding this comment.
Avoid describing the repo as a single “插件”.
This repository is presented here as a collection of agents, skills, commands, hooks, rules, and MCP configs, so calling it a single Claude Code plugin is misleading.
✏️ Suggested wording
-这是一个 **Claude Code 插件** - 一个包含生产就绪的代理、技能、钩子、命令、规则和 MCP 配置的集合。该项目提供了使用 Claude Code 进行软件开发的经验证的工作流。
+这是一个 **Claude Code 资源集合**，包含生产就绪的代理、技能、钩子、命令、规则和 MCP 配置。该项目提供了使用 Claude Code 进行软件开发的经验证的工作流。📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| 这是一个 **Claude Code 插件** - 一个包含生产就绪的代理、技能、钩子、命令、规则和 MCP 配置的集合。该项目提供了使用 Claude Code 进行软件开发的经验证的工作流。 | |
| 这是一个 **Claude Code 资源集合**，包含生产就绪的代理、技能、钩子、命令、规则和 MCP 配置。该项目提供了使用 Claude Code 进行软件开发的经验证的工作流。 |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/CLAUDE.md` at line 7, The sentence that calls the repo "这是一个
**Claude Code 插件**" is misleading; update the opening line in CLAUDE.md to
describe the repository as a collection of components (agents, skills, hooks,
commands, rules, and MCP configs) rather than a single plugin—replace the phrase
"Claude Code 插件" with wording such as "一组 Claude Code 组件/资源" or
"包含代理、技能、钩子、命令、规则和 MCP 配置的集合" and ensure the rest of the sentence still conveys
the production-ready workflows for using Claude Code.
Comment on lines
+39
to
+40
| 辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告，邮箱为。 | ||
| 所有投诉都将得到及时和公正的审查和调查。 |
Contributor
There was a problem hiding this comment.
Add an enforcement contact.
The reporting path is incomplete here: the text says complaints can be sent by email, but no address is provided. As written, contributors have no way to report Code of Conduct violations.
🧰 Tools
🪛 LanguageTool
[uncategorized] ~40-~40: 动词的修饰一般为‘形容词（副词）+地+动词’。您的意思是否是：公正"地"审查
Context: ...受行为的实例可以向负责执行的社区领导者报告，邮箱为。 所有投诉都将得到及时和公正的审查和调查。 所有社区领导者都有义务尊重任何事件报告者的隐私和安全。 ## ...
(wb4)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/CODE_OF_CONDUCT.md` around lines 39 - 40, The enforcement contact
is missing in the sentence that currently reads
"辱骂、骚扰或其他不可接受行为的实例可以向负责执行的社区领导者报告，邮箱为。"; update that line in CODE_OF_CONDUCT.md
to include a valid enforcement email address or a placeholder (e.g.,
[email protected] or [email protected]) so contributors know where to send
reports, and ensure the updated sentence keeps the same wording and punctuation
around the added address.
Comment on lines
+22
to
23
| /instinct-export --scope project --output project-instincts.yaml | ||
| ``` |
Contributor
There was a problem hiding this comment.
Call out the export/import scope mismatch explicitly.
/instinct-export now defaults to --scope all, but /instinct-import still defaults to project. A user following these docs for backup/share/restore can silently re-import global instincts as project-scoped unless the import side is documented with an explicit scope. Please add a note or paired example so the round-trip preserves intent.
Also applies to: 27-33, 64-67
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/commands/instinct-export.md` around lines 22 - 23, The docs show
an /instinct-export example but don’t call out that export now defaults to
--scope all while /instinct-import still defaults to project, so add an explicit
note and matching import example that sets the same --scope (e.g., use --scope
project in both export and import or --scope all in both) to preserve intent;
update the example block with /instinct-import and add a brief warning sentence
about the default-scope mismatch and mirror this change in the other example
ranges referenced (the example pairs around the blocks currently at lines 27-33
and 64-67) so readers see a round-trip that preserves scope.
Comment on lines
+615
to
+623
| # Copy skills (core vs niche) | ||
| # Recommended (new users): core/general skills only | ||
| cp -r everything-claude-code/.agents/skills/* ~/.claude/skills/ | ||
| cp -r everything-claude-code/skills/search-first ~/.claude/skills/ | ||
|
|
||
| # Optional: add niche/framework-specific skills only when needed | ||
| # for s in django-patterns django-tdd springboot-patterns; do | ||
| # cp -r everything-claude-code/skills/$s ~/.claude/skills/ | ||
| # done |
Contributor
There was a problem hiding this comment.
Manual install now points Claude Code users at the wrong skills directory.
This section switched to everything-claude-code/.agents/skills/*, but the repo tree above documents Claude Code skills under skills/, and .agents/skills/ is only introduced later for Codex. Copy-pasting these commands will either fail on a normal checkout or install the wrong skill format.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/README.md` around lines 615 - 623, The manual install commands in
README.md currently copy from everything-claude-code/.agents/skills/* which is
incorrect for Claude Code (the repo documents Claude Code skills under skills/);
update the copy commands to use everything-claude-code/skills/ (and keep the
separate copy for everything-claude-code/skills/search-first) so users copy the
correct Claude Code skill format instead of the Codex .agents/skills format;
update or add a brief note near the cp commands clarifying that .agents/skills/
is for Codex only if you want to keep that alternative mentioned.
Comment on lines
+72
to
+82
| 默认为 **核心（推荐给新用户）** — 对于研究优先的工作流，复制 `.agents/skills/*` 加上 `skills/search-first/`。此捆绑包涵盖工程、评估、验证、安全、战略压缩、前端设计以及 Anthropic 跨职能技能（文章写作、内容引擎、市场研究、前端幻灯片）。 | ||
|
|
||
| 使用 `AskUserQuestion`（单选）： | ||
|
|
||
| ``` | ||
| Question: "Install core skills only, or include niche/framework packs?" | ||
| Options: | ||
| - "Core only (recommended)" — "tdd, e2e, evals, verification, research-first, security, frontend patterns, compacting, cross-functional Anthropic skills" | ||
| - "Core + selected niche" — "Add framework/domain-specific skills after core" | ||
| - "Niche only" — "Skip core, install specific framework/domain skills" | ||
| Default: Core only |
Contributor
There was a problem hiding this comment.
Keep the core preset on canonical skill paths and IDs.
This block switches to .agents/skills/* and summary labels like research-first / evals, but the install flow below and the canonical English source use concrete directories under $ECC_ROOT/skills/<skill-name>. As written, the Chinese doc no longer names a stable set of installable skills for the “core” preset.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/skills/configure-ecc/SKILL.md` around lines 72 - 82, The Chinese
doc replaced canonical skill paths/IDs with generic labels (e.g.,
`.agents/skills/*`, `research-first`, `evals`) causing the "core" preset to be
ambiguous; update the block that defines the core preset and the AskUserQuestion
options to use the concrete, canonical skill directories and IDs from the
English source (use `$ECC_ROOT/skills/<skill-name>` style names) so the install
flow maps to stable installable skills (e.g., list the actual skill folder names
under `$ECC_ROOT/skills/` instead of labels like `research-first`), and ensure
the AskUserQuestion options and Default still reference the canonical "Core only
(recommended)" set of skill IDs.
Comment on lines
+201
to
+243
| ## ECC v1.8 新增内容 | ||
|
|
||
| ### 可复制的钩子配置文件 | ||
|
|
||
| 设置严格的质量行为： | ||
|
|
||
| ```bash | ||
| export ECC_HOOK_PROFILE=strict | ||
| export ECC_QUALITY_GATE_FIX=true | ||
| export ECC_QUALITY_GATE_STRICT=true | ||
| ``` | ||
|
|
||
| ### 语言关卡表 | ||
|
|
||
| * TypeScript/JavaScript：首选 Biome，Prettier 作为后备 | ||
| * Python：Ruff 格式/检查 | ||
| * Go：gofmt | ||
|
|
||
| ### 配置篡改防护 | ||
|
|
||
| 在质量强制执行期间，标记同一迭代中对配置文件的更改： | ||
|
|
||
| * `biome.json`, `.eslintrc*`, `prettier.config*`, `tsconfig.json`, `pyproject.toml` | ||
|
|
||
| 如果配置被更改以抑制违规，则要求在合并前进行明确审查。 | ||
|
|
||
| ### CI 集成模式 | ||
|
|
||
| 在 CI 中使用与本地钩子相同的命令： | ||
|
|
||
| 1. 运行格式化程序检查 | ||
| 2. 运行 lint/类型检查 | ||
| 3. 严格模式下快速失败 | ||
| 4. 发布修复摘要 | ||
|
|
||
| ### 健康指标 | ||
|
|
||
| 跟踪： | ||
|
|
||
| * 被关卡标记的编辑 | ||
| * 平均修复时间 | ||
| * 按类别重复违规 | ||
| * 因关卡失败导致的合并阻塞 |
Contributor
There was a problem hiding this comment.
Clarify ECC vs Plankton feature attribution.
This section titled "ECC v1.8 additions" appears in Plankton skill documentation and creates confusion about which features belong to ECC versus Plankton:
-
Lines 213-217 "Language gate table": Describes Biome/Prettier/Ruff/gofmt approach that matches ECC's actual quality-gate.js implementation (per context snippet 3), not Plankton's comprehensive multi-linter system described earlier.
-
Mixed environment variables: Shows ECC-specific env vars (
ECC_HOOK_PROFILE,ECC_QUALITY_GATE_FIX) in a Plankton document without clearly explaining the relationship. -
Config tampering protection (lines 219-225): Lists files that differ from Plankton's protection list described in lines 60-62, creating inconsistency.
-
Unclear scope: The "CI integration mode" and "health metrics" subsections don't specify whether they're ECC features, Plankton features, or integration patterns.
Recommendation: Either:
- Remove this section if it's not relevant to Plankton
- Rename to "Integration with ECC v1.8" and clearly distinguish ECC features from Plankton features
- Move ECC-specific content to the "Pairing with ECC" section (lines 116-142) where ECC/Plankton differences are already explained
This is critical for preventing user confusion about which project provides which capabilities.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/skills/plankton-code-quality/SKILL.md` around lines 201 - 243,
This "ECC v1.8 新增内容" block mixes ECC-specific features into the Plankton skill
doc; either remove it or clearly mark it as ECC integration. Fix by moving the
entire section into the existing "Pairing with ECC" area (refer to the "Pairing
with ECC" section), or rename the header to "Integration with ECC v1.8" and
annotate each subsection (e.g., "Language gate table", the env vars
ECC_HOOK_PROFILE/ECC_QUALITY_GATE_FIX/ECC_QUALITY_GATE_STRICT, "配置篡改防护", "CI
集成模式", "健康指标") to explicitly state these are ECC features and reconcile the
listed config files with Plankton's protection list.
| *** | ||
|
|
||
| *基于 [Supabase Agent Skills](Supabase Agent Skills (credit: Supabase team)) (MIT License)* | ||
| *基于 Supabase 代理技能（致谢：Supabase 团队）（MIT 许可证）* |
Contributor
There was a problem hiding this comment.
Keep the upstream source link in the attribution.
The Chinese wording is fine, but removing the original reference makes provenance and license verification harder for future maintainers/readers. Please keep the localized text and restore the source link.
💡 Suggested change
-*基于 Supabase 代理技能（致谢：Supabase 团队）（MIT 许可证）*
+*基于 [Supabase 代理技能](Supabase Agent Skills (credit: Supabase team))（致谢：Supabase 团队，MIT 许可证）*🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/skills/postgres-patterns/SKILL.md` at line 154, Restore the
upstream source link in the attribution line by editing the string '*基于 Supabase
代理技能（致谢：Supabase 团队）（MIT 许可证）*' to include the original upstream reference URL
(e.g. the Supabase repo or source file link) while keeping the Chinese
localization intact; ensure the final line clearly contains both the localized
attribution and the original source link so provenance and license verification
remain visible.
Comment on lines
+139
to
+140
|  | ||
| *在我们的 50 个任务基准测试中，mgrep + Claude Code 在相似或更好的判断质量下，使用的 token 数比基于 grep 的工作流少约 2 倍。来源：@mixedbread-ai 的 mgrep* |
Contributor
There was a problem hiding this comment.
Polish the benchmark sentence for natural Chinese.
Line 140 reads awkwardly as written, and “少约 2 倍” is ambiguous. A more direct phrasing will make the claim easier to understand.
✏️ Suggested wording
-
-*在我们的 50 个任务基准测试中，mgrep + Claude Code 在相似或更好的判断质量下，使用的 token 数比基于 grep 的工作流少约 2 倍。来源：@mixedbread-ai 的 mgrep*
+
+*在我们的 50 个任务基准测试中，mgrep + Claude Code 在判断质量相近或更高的情况下，所用 token 数量约为基于 grep 工作流的一半。来源：@mixedbread-ai 的 mgrep*📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
|  | |
| *在我们的 50 个任务基准测试中，mgrep + Claude Code 在相似或更好的判断质量下，使用的 token 数比基于 grep 的工作流少约 2 倍。来源：@mixedbread-ai 的 mgrep* | |
|  | |
| *在我们的 50 个任务基准测试中，mgrep + Claude Code 在判断质量相近或更高的情况下，所用 token 数量约为基于 grep 工作流的一半。来源：@mixedbread-ai 的 mgrep* |
🧰 Tools
🪛 LanguageTool
[uncategorized] ~140-~140: 动词的修饰一般为‘形容词（副词）+地+动词’。您的意思是否是：好"地"判断
Context: ...的 50 个任务基准测试中，mgrep + Claude Code 在相似或更好的判断质量下，使用的 token 数比基于 grep 的工作流少约 2 倍。来源：...
(wb4)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/zh-CN/the-longform-guide.md` around lines 139 - 140, Replace the awkward
Chinese sentence in the image caption describing the mgrep benchmark so it reads
naturally and unambiguously: update the sentence mentioning "mgrep + Claude
Code" and the token usage comparison to state that it uses about half the tokens
of the grep-based workflow (e.g., "使用的 token 约为基于 grep 的工作流的一半" or "节省了约一半的
token") and keep the source attribution "来源：@mixedbread-ai 的 mgrep"; ensure the
caption remains a single fluent sentence in Chinese.
Contributor
There was a problem hiding this comment.
♻️ Duplicate comments (2)
docs/zh-CN/README.md (2)
615-623:⚠️ Potential issue | 🟠 MajorManual install commands point to wrong skills directory for Claude Code.
Line 617 copies from
everything-claude-code/.agents/skills/*, but the repo documents Claude Code skills underskills/(.agents/skills/is for Codex). Users following these instructions will install the wrong skill format or encounter missing files. This was flagged in a previous review and remains unresolved.📝 Suggested fix
# Copy skills (core vs niche) # Recommended (new users): core/general skills only -cp -r everything-claude-code/.agents/skills/* ~/.claude/skills/ +cp -r everything-claude-code/skills/* ~/.claude/skills/ cp -r everything-claude-code/skills/search-first ~/.claude/skills/Or clarify that line 617 is specifically for Codex users:
# Copy skills (core vs niche) -# Recommended (new users): core/general skills only -cp -r everything-claude-code/.agents/skills/* ~/.claude/skills/ +# For Claude Code: copy from skills/ +cp -r everything-claude-code/skills/* ~/.claude/skills/ +# For Codex: copy from .agents/skills/ +# cp -r everything-claude-code/.agents/skills/* ~/.claude/skills/ cp -r everything-claude-code/skills/search-first ~/.claude/skills/🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/README.md` around lines 615 - 623, The manual install commands copy skills from the wrong directory — update the cp commands that reference everything-claude-code/.agents/skills/* to use everything-claude-code/skills/* (and ensure the search-first line uses everything-claude-code/skills/search-first) so users install Claude Code skill files (not Codex .agents/skills); alternatively, add a clarifying note above these commands that .agents/skills is for Codex only and show separate commands for Codex vs Claude Code so the intended skill directories (skills/ vs .agents/skills/) are unambiguous.
10-10:⚠️ Potential issue | 🟡 MinorStatic install count badge will become outdated.
The GitHub App badge hardcodes "150 installs" which will drift as adoption grows. This was flagged in a previous review and remains unresolved.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@docs/zh-CN/README.md` at line 10, The README currently contains a hardcoded GitHub App install badge (the markdown starting with "[](https://github.com/marketplace/ecc-tools)"); replace it so the install count is not static — either remove the hardcoded "150 installs" badge or replace it with a dynamic badge (e.g., a shields.io or GitHub Marketplace dynamic badge URL that reflects live install counts) and keep the link target to the marketplace page.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Duplicate comments:
In `@docs/zh-CN/README.md`:
- Around line 615-623: The manual install commands copy skills from the wrong
directory — update the cp commands that reference
everything-claude-code/.agents/skills/* to use everything-claude-code/skills/*
(and ensure the search-first line uses
everything-claude-code/skills/search-first) so users install Claude Code skill
files (not Codex .agents/skills); alternatively, add a clarifying note above
these commands that .agents/skills is for Codex only and show separate commands
for Codex vs Claude Code so the intended skill directories (skills/ vs
.agents/skills/) are unambiguous.
- Line 10: The README currently contains a hardcoded GitHub App install badge
(the markdown starting with "[](https://github.com/marketplace/ecc-tools)");
replace it so the install count is not static — either remove the hardcoded "150
installs" badge or replace it with a dynamic badge (e.g., a shields.io or GitHub
Marketplace dynamic badge URL that reflects live install counts) and keep the
link target to the marketplace page.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Type of Change
fix:Bug fixfeat:New featurerefactor:Code refactoringdocs:Documentationtest:Testschore:Maintenance/toolingci:CI/CD changesChecklist
node tests/run-all.js)Summary by cubic
Synced zh-CN docs to upstream v1.8, adding new agents, commands, and skills, and updating guides for NanoClaw v2, hook reliability, and continuous-learning v2.1. Also fixed README language links and clarified marketplace/sponsor references.
New Features
Migration
Written for commit 6f4b603. Summary will update on new commits.
Summary by CodeRabbit
New Features
Documentation