feat(skills): add polishpro academic research skill

skills/polishpro/SKILL.md

@@ -0,0 +1,62 @@
+---
+name: polishpro
+version: 1.0.0
+description: "学术英文润色检查器：对英文草稿做句级检查，标出冗余短语（in order to→to）、弱动词/名词化（make an investigation of→investigate）、被动语态滥用、中式英文/赘词（very important、more and more），逐条给出更简洁地道的改法建议。当用户提到「英文润色」「检查英文表达」「中式英文」「句子啰嗦」「英文不地道」「polish」时触发。"
+metadata:
+  requires:
+    optional_bins: ["python"]
+---
+
+# polishpro（学术英文润色检查器）
+
+英文写得出、但总被说「表达不地道」？polishpro 不改错字，而是把审稿人读着别扭的
+表达——啰嗦、中式、被动滥用——逐句标出来，给更简洁地道的改法。
+
+## 何时使用本技能
+
+- 自己写的英文怕不地道，过一遍
+- 句子啰嗦，想压简洁
+- 投稿前做一次表达自查
+
+## 与同类工具的区别
+
+同类润色工具多直接调 LLM 重写。本技能是**纯规则引擎**：不联网、不调 LLM，基于学术
+写作公认的简洁性原则逐条检查，给的是**带解释的改进建议清单**而非黑盒重写。它的价值
+在「指出问题 + 教你为什么」，最终改不改、怎么改由你定——既可控、可离线，也帮你
+积累地道表达的直觉。
+
+## 四类检查
+
+| 类别 | 例子 |
+|------|------|
+| 冗余短语 | in order to→to，due to the fact that→because |
+| 弱动词/名词化 | make an investigation of→investigate |
+| 被动语态 | 标出被动结构，提示非方法章节可改主动 |
+| 中式英文/赘词 | very important、more and more→increasingly |
+
+另给平均句长与过长句子（>35 词）提示，建议拆分。
+
+## 工作流
+
+```bash
+python scripts/polish.py --input draft.txt --output report.md
+python scripts/polish.py --text "In order to improve..." --format json
+```
+
+| 参数 | 说明 |
+|------|------|
+| `--input` | 英文草稿文件（*或用 `--text`） |
+| `--text` | 直接传入英文文本 |
+| `--format` | `markdown`（默认）或 `json` |
+| `--output` | 输出文件 |
+
+## 注意事项
+
+- 规则覆盖学术写作高频问题，但非穷尽；最终判断结合语境。
+- 被动语态在方法章节是规范的，提示仅供参考，不必一律改主动。
+- 工具只给建议，不替你重写句子——避免改变你的原意。
+
+## References
+
+- [rules.md](references/rules.md) — 四类检查的规则与建议对照
+- [usage.md](references/usage.md) — 使用建议与配合写作流程

skills/polishpro/references/rules.md

@@ -0,0 +1,56 @@
+# 四类检查规则与建议对照
+
+polishpro 基于学术写作公认的简洁性原则，逐句匹配以下规则。规则定义在 `polish.py`
+的四个字典/正则中，可按需扩充。
+
+## 冗余短语（WORDY_PHRASES）
+
+| 啰嗦写法 | 建议 |
+|---------|------|
+| in order to | to |
+| due to the fact that | because |
+| in spite of the fact that | although |
+| a number of | many / few |
+| in the event that | if |
+| it is worth noting that | notably / 直接陈述 |
+| with regard to | regarding / about |
+| for the purpose of | to / for |
+| as a matter of fact | 删去 |
+
+## 弱动词 / 名词化（NOMINALIZATIONS）
+
+把「动词被名词化再配弱动词（make/conduct/perform…）」还原成强动词：
+
+| 啰嗦 | 强动词 |
+|------|--------|
+| make an investigation of | investigate |
+| conduct an analysis of | analyze |
+| perform a comparison of | compare |
+| give an explanation of | explain |
+| make use of | use |
+
+## 被动语态（PASSIVE）
+
+正则匹配「be 动词 + 过去分词」。被动在方法章节是规范的（强调操作而非操作者），
+但在引言/讨论中过多被动会让表达迂回。工具只标出、不强制改。
+
+## 中式英文 / 赘词（CHINGLISH）
+
+| 表达 | 建议 |
+|------|------|
+| more better | better |
+| very important | 去 very，用精确词 |
+| more and more | increasingly |
+| as we all know | 学术写作避免 |
+| big difference | large / substantial |
+| get better | improve |
+
+## 句长
+
+平均句长与单句词数：超过 35 词的句子会被标为「过长，建议拆分」。学术英文并非越长
+越严谨，长句易丢失主干。
+
+## 设计原则
+
+工具给「建议」不做「重写」——保留你的原意，只指出问题并解释为什么。这样你改的同时
+也在积累地道表达的判断力。

skills/polishpro/references/usage.md

@@ -0,0 +1,29 @@
+# 使用建议与写作流程
+
+## 怎么用
+
+1. 把英文草稿（段落或全文）存成 .txt，或用 `--text` 直接传。
+2. 运行得到检查报告：分类统计 + 逐条建议 + 过长句提示。
+3. 对照建议逐条改，改不改由你定——建议都附了「为什么」。
+
+## 报告怎么读
+
+- **分类统计**：先看哪类问题最多，集中优化。冗余短语和被动语态通常最常见。
+- **逐条建议**：每条给出原片段、问题类型、建议改法、所在整句，方便定位。
+- **过长句子**：>35 词的句子单列，优先考虑拆分。
+
+## 配合写作流程
+
+- draftsmith 起草 → 自己填实内容 → **polishpro 过一遍英文** → 投稿。
+- polishpro 检查的是「表达」，不检查「事实/逻辑」，后者靠 peermirror 自审。
+
+## 注意
+
+- 被动语态提示要结合章节：方法章节用被动是规范的，不必强行改主动。
+- 规则是高频问题的覆盖，不是穷尽；它帮你抓大部分明显问题，细腻语感仍需人来把关。
+- 工具不重写句子，避免改变原意——它是「教练」不是「代笔」。
+
+## 局限
+
+- 基于规则与正则，无法理解语义，可能漏掉「语法对但表达绕」的句子。
+- 仅针对英文；中英混排文本的非英文部分会被忽略。

skills/polishpro/scripts/polish.py

@@ -0,0 +1,209 @@
+"""polishpro：学术英文润色检查器。
+
+对英文草稿做句级检查，把「中式英文 / 啰嗦 / 不地道」的地方标出来，给出更简洁、
+更地道的改法建议。它不是改错字，而是把审稿人读着别扭的表达，指出来并建议怎么改。
+
+四类检查（纯规则，不调 LLM）：
+1. 冗余短语——in order to→to、due to the fact that→because、a number of→many 等
+   学术写作里高频的啰嗦表达，建议精简。
+2. 弱动词 / 名词化——make an investigation of→investigate、conduct an analysis→
+   analyze，把「动词被名词化再配弱动词」的结构还原成强动词。
+3. 被动语态提示——标出被动结构，提示在方法之外的章节可考虑主动语态。
+4. 中式英文 / 赘词——more better、according to、very important 等中式或口语化表达。
+
+每条命中给出：原片段、问题类型、建议改法、所在句子。纯本地、纯标准库，离线可用。
+它给的是「改进建议清单」，最终改不改、怎么改由你定。
+
+用法：
+    python polish.py --input draft.txt --output report.md
+    python polish.py --text "In order to improve the performance..." --format json
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+import re
+import sys
+from pathlib import Path
+from typing import Any
+
+if hasattr(sys.stdout, "reconfigure"):
+    try:
+        sys.stdout.reconfigure(encoding="utf-8")
+    except Exception:
+        pass
+
+
+# 冗余短语：模式 -> 建议
+WORDY_PHRASES = {
+    r"\bin order to\b": "to",
+    r"\bdue to the fact that\b": "because",
+    r"\bin spite of the fact that\b": "although",
+    r"\ba (large |small )?number of\b": "many / few",
+    r"\bin the event that\b": "if",
+    r"\bat this point in time\b": "now",
+    r"\bin the near future\b": "soon",
+    r"\bit is worth noting that\b": "notably / （直接陈述）",
+    r"\bit should be noted that\b": "note that / （直接陈述）",
+    r"\bwith regard to\b": "regarding / about",
+    r"\bin terms of\b": "（常可删去或换具体说法）",
+    r"\bfor the purpose of\b": "to / for",
+    r"\bin the case of\b": "for / in",
+    r"\bas a matter of fact\b": "（多余，建议删）",
+}
+
+# 名词化 + 弱动词 -> 强动词
+NOMINALIZATIONS = {
+    r"\bmake an investigation of\b": "investigate",
+    r"\bconduct an analysis of\b": "analyze",
+    r"\bperform a comparison of\b": "compare",
+    r"\bgive an explanation of\b": "explain",
+    r"\bprovide a description of\b": "describe",
+    r"\bcarry out an evaluation of\b": "evaluate",
+    r"\bmake use of\b": "use",
+    r"\bmake a decision\b": "decide",
+}
+
+# 中式英文 / 赘词 / 不地道
+CHINGLISH = {
+    r"\bmore better\b": "better（more 多余）",
+    r"\bvery (important|significant|crucial)\b": "（去掉 very，用更精确的词或直接陈述）",
+    r"\bmore and more\b": "increasingly",
+    r"\bas we all know\b": "（学术写作避免，直接陈述或引用）",
+    r"\bin a word\b": "in short / in summary",
+    r"\bby the way\b": "（学术写作不用）",
+    r"\bso\b": "（句首 so 偏口语，用 therefore/thus）",
+    r"\bbig (difference|improvement|effect)\b": "large / substantial",
+    r"\bget (better|worse)\b": "improve / deteriorate",
+}
+
+# 被动语态（粗检：be 动词 + 过去分词）
+PASSIVE = re.compile(
+    r"\b(is|are|was|were|be|been|being)\s+(\w+ed|done|made|shown|found|used|given|taken|seen|known)\b",
+    re.IGNORECASE)
+
+
+def split_sentences(text: str) -> list[str]:
+    text = text.lstrip("\ufeff")
+    text = re.sub(r"\s+", " ", text.strip())
+    # 简单断句（润色场景对断句精度要求不高）
+    parts = re.split(r"(?<=[.!?])\s+(?=[A-Z])", text)
+    return [p.strip() for p in parts if p.strip()]
+
+
+def _check_dict(sentence: str, patterns: dict[str, str], category: str) -> list[dict[str, Any]]:
+    issues = []
+    for pat, suggestion in patterns.items():
+        for m in re.finditer(pat, sentence, re.IGNORECASE):
+            issues.append({
+                "category": category,
+                "fragment": m.group(0),
+                "suggestion": suggestion,
+                "sentence": sentence,
+            })
+    return issues
+
+
+def check_sentence(sentence: str) -> list[dict[str, Any]]:
+    issues: list[dict[str, Any]] = []
+    issues += _check_dict(sentence, WORDY_PHRASES, "冗余短语")
+    issues += _check_dict(sentence, NOMINALIZATIONS, "弱动词/名词化")
+    issues += _check_dict(sentence, CHINGLISH, "中式英文/赘词")
+    # 被动语态
+    for m in PASSIVE.finditer(sentence):
+        issues.append({
+            "category": "被动语态",
+            "fragment": m.group(0),
+            "suggestion": "方法章节可保留；其余章节考虑改主动语态使表达更直接",
+            "sentence": sentence,
+        })
+    return issues
+
+
+def polish(text: str) -> dict[str, Any]:
+    sentences = split_sentences(text)
+    all_issues: list[dict[str, Any]] = []
+    for s in sentences:
+        all_issues.extend(check_sentence(s))
+    by_cat: dict[str, int] = {}
+    for it in all_issues:
+        by_cat[it["category"]] = by_cat.get(it["category"], 0) + 1
+    # 简单可读性提示：平均句长
+    words = sum(len(re.findall(r"\b\w+\b", s)) for s in sentences)
+    avg_len = round(words / len(sentences), 1) if sentences else 0
+    long_sents = [s for s in sentences if len(re.findall(r"\b\w+\b", s)) > 35]
+    return {
+        "sentence_count": len(sentences),
+        "issue_count": len(all_issues),
+        "by_category": by_cat,
+        "issues": all_issues,
+        "avg_sentence_length": avg_len,
+        "long_sentences": long_sents,
+    }
+
+
+def render_markdown(data: dict[str, Any]) -> str:
+    lines = [
+        "# 英文润色检查报告",
+        "",
+        f"共 {data['sentence_count']} 句，发现 {data['issue_count']} 处可改进，"
+        f"平均句长 {data['avg_sentence_length']} 词。",
+        "",
+        "## 分类统计",
+        "",
+    ]
+    if data["by_category"]:
+        for cat, n in sorted(data["by_category"].items(), key=lambda kv: -kv[1]):
+            lines.append(f"- {cat}：{n} 处")
+    else:
+        lines.append("- 未发现明显问题，表达较简洁。")
+    lines += ["", "## 逐条建议", ""]
+    if data["issues"]:
+        for i, it in enumerate(data["issues"], start=1):
+            lines.append(f"{i}. [{it['category']}] `{it['fragment']}` → 建议：{it['suggestion']}")
+            lines.append(f"   > 所在句：{it['sentence']}")
+            lines.append("")
+    else:
+        lines.append("（无）")
+        lines.append("")
+    if data["long_sentences"]:
+        lines += ["## 过长句子（>35 词，建议拆分）", ""]
+        for s in data["long_sentences"]:
+            lines.append(f"- {s}")
+        lines.append("")
+    lines += ["---", "",
+              "由 polishpro 生成。基于规则的改进建议，最终改法由你定；它不替你重写句子。"]
+    return "\n".join(lines)
+
+
+def main(argv: list[str] | None = None) -> int:
+    p = argparse.ArgumentParser(prog="polishpro", description="学术英文润色检查器")
+    p.add_argument("--input", type=Path, help="英文草稿文件")
+    p.add_argument("--text", help="直接传入英文文本")
+    p.add_argument("--format", choices=["markdown", "json"], default="markdown")
+    p.add_argument("--output", type=Path)
+    args = p.parse_args(argv)
+
+    if args.input and args.input.exists():
+        text = args.input.read_text(encoding="utf-8-sig", errors="replace")
+    elif args.text:
+        text = args.text
+    else:
+        print("错误：请用 --input 或 --text 提供英文文本。", file=sys.stderr)
+        return 2
+
+    data = polish(text)
+    out = (json.dumps(data, ensure_ascii=False, indent=2) if args.format == "json"
+           else render_markdown(data))
+    if args.output:
+        args.output.parent.mkdir(parents=True, exist_ok=True)
+        args.output.write_text(out, encoding="utf-8")
+        print(f"已写入 {args.output}")
+    else:
+        print(out)
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

skills/polishpro/tests/test_polish.py

@@ -0,0 +1,75 @@
+"""polishpro 单元测试。"""
+from __future__ import annotations
+import sys
+from pathlib import Path
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "scripts"))
+import pytest
+from polish import split_sentences, check_sentence, polish, render_markdown
+
+
+class TestSplit:
+    def test_basic(self):
+        assert len(split_sentences("First one. Second one. Third one.")) == 3
+
+    def test_bom(self):
+        s = split_sentences("\ufeffHello world. Bye.")
+        assert not s[0].startswith("\ufeff")
+
+
+class TestCheckSentence:
+    def test_wordy(self):
+        issues = check_sentence("In order to test it.")
+        assert any(i["category"] == "冗余短语" and "in order to" in i["fragment"].lower() for i in issues)
+
+    def test_nominalization(self):
+        issues = check_sentence("We make an investigation of the data.")
+        cats = [i["category"] for i in issues]
+        assert "弱动词/名词化" in cats
+
+    def test_chinglish(self):
+        issues = check_sentence("This is a very important result.")
+        assert any(i["category"] == "中式英文/赘词" for i in issues)
+
+    def test_passive(self):
+        issues = check_sentence("The model was trained on the dataset.")
+        assert any(i["category"] == "被动语态" for i in issues)
+
+    def test_clean_sentence(self):
+        issues = check_sentence("We trained the model.")
+        # 主动、简洁，应无或极少问题
+        assert all(i["category"] != "冗余短语" for i in issues)
+
+
+class TestPolish:
+    def test_counts(self):
+        data = polish("In order to win, we make use of tricks. Due to the fact that it works.")
+        assert data["issue_count"] >= 3
+        assert data["sentence_count"] == 2
+
+    def test_by_category(self):
+        data = polish("In order to improve, due to the fact that data is limited.")
+        assert "冗余短语" in data["by_category"]
+
+    def test_avg_length(self):
+        data = polish("Short one. Another short.")
+        assert data["avg_sentence_length"] > 0
+
+    def test_long_sentence_flagged(self):
+        long = " ".join(["word"] * 40) + "."
+        data = polish(long)
+        assert data["long_sentences"]
+
+
+class TestRender:
+    def test_render_with_issues(self):
+        md = render_markdown(polish("In order to test, we make use of data."))
+        assert "英文润色检查报告" in md
+        assert "逐条建议" in md
+
+    def test_render_clean(self):
+        md = render_markdown(polish("We trained the model."))
+        assert "英文润色检查报告" in md
+
+
+if __name__ == "__main__":
+    sys.exit(pytest.main([__file__, "-v"]))