duyunzhoudyz/dyz-gongwen-skills / skills/gongwen-review
gongwen-review
Verified按“八关”审核公文和材料,输出分级问题清单和批注用 JSON,含机检、提法风险、保密隐私扫描、本单位禁用词比对;默认只提意见,用户说“直接改”才按 P0、P1 改。用于“帮我核一下稿”“把把关”“送签前过一遍”“看看报告里是不是夹带请示”“查查格式标点”“审一下报来的材料”。大幅改写转 gongwen-polish 或原起草 skill;去 AI 腔转 gongwen-humanize;去免责对冲转 gongwen-direct;排版、写批注转 gongwen-format;信息能不能被采用的评估交 gongwen-brief;装有行业包时,本行业专用文书只查行业问题的交对应行业包。
scripts/secret_check.py
View on GitHub#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""保密标识检查:扫描 txt、md、docx、xlsx 中的密级标识和内部标注,命中即退出码 1。
用法:python secret_check.py <文件...>
docx 读取范围:正文段落、表格、页眉、页脚(含首页、奇偶页)、文本框、脚注、尾注、批注,
输出位置带来源标签,如“[页眉(首页)]”“[文本框]”“[表格]”。xlsx 读取单元格文字、批注、
工作表名、页眉页脚。只依赖标准库。
检出:“秘密★”“机密★”“绝密★”(含 5年 等保密期限写法)、“工作秘密”“内部资料”
“不得外传”“内部使用”等。“保密工作”“涉密人员管理”之类业务用语不算。
输出只给位置和命中词,不复述材料内容。
退出码:0 未发现;1 发现标识;2 文件无法读取。
"""
from __future__ import annotations
import argparse
import re
import sys
import zipfile
import xml.etree.ElementTree as ET
from pathlib import Path
from typing import Dict, List, Tuple
W = "{http://schemas.openxmlformats.org/wordprocessingml/2006/main}"
R = "{http://schemas.openxmlformats.org/officeDocument/2006/relationships}"
S = "{http://schemas.openxmlformats.org/spreadsheetml/2006/main}"
MAX_XML = 50 * 1024 * 1024
LEVEL = "绝密|机密|秘密"
STAR = "[★☆**]"
PATTERNS = [
("密级标识", re.compile("(?:%s)\\s*%s" % (LEVEL, STAR))),
("密级与保密期限", re.compile("(?:%s)\\s*%s?\\s*[\\d一二三四五六七八九十]+\\s*(?:年|个月)" % (LEVEL, STAR))),
("密级与保密期限", re.compile("(?:%s)\\s*%s?\\s*(?:长期|永久)" % (LEVEL, STAR))),
("保密期限", re.compile("保密期限")),
("内部标注", re.compile("工作秘密|内部资料|不得外传|严禁外传|内部使用|仅供内部参考|内部参阅|注意保密")),
]
Item = Tuple[str, str] # (位置标签, 文字)
def _root(zf: zipfile.ZipFile, name: str) -> ET.Element:
if zf.getinfo(name).file_size > MAX_XML:
raise ValueError("部件过大:" + name)
return ET.fromstring(zf.read(name))
def _para_text(p: ET.Element) -> str:
"""段落文字:含修订删除的文字(保守起见一并检查)。"""
parts = []
for e in p.iter():
if e.tag in (W + "t", W + "delText"):
parts.append(e.text or "")
elif e.tag == W + "tab":
parts.append("\t")
return "".join(parts)
def _paragraphs(root: ET.Element, base: str) -> List[Item]:
"""遍历部件内全部段落;文本框、表格内的段落单独标注来源。"""
out: List[Item] = []
counter = [0]
def walk(el: ET.Element, ctx: str) -> None:
for ch in el:
if ch.tag == W + "txbxContent":
walk(ch, "文本框")
elif ch.tag == W + "tbl":
walk(ch, ctx if ctx == "文本框" else "表格")
elif ch.tag == W + "p":
counter[0] += 1
text = _para_text(ch)
if text.strip():
if ctx == "正文":
label = "%s第%d段" % (base, counter[0]) if base == "正文" else base
else:
label = ctx if base == "正文" else base + ctx
out.append((label, text))
walk(ch, ctx)
else:
walk(ch, ctx)
walk(root, "正文")
seen = set() # mc:AlternateContent 中文本框常有两份相同内容,去重
uniq = []
for item in out:
if item not in seen:
seen.add(item)
uniq.append(item)
return uniq
def _hf_labels(zf: zipfile.ZipFile) -> Dict[str, str]:
"""部件路径 -> 页眉/页脚标签(区分首页、偶数页)。"""
labels: Dict[str, str] = {}
names = set(zf.namelist())
if "word/document.xml" not in names:
return labels
rels = {}
rp = "word/_rels/document.xml.rels"
if rp in names:
for r in _root(zf, rp):
rels[r.get("Id")] = r.get("Target", "")
kinds = {"default": "", "first": "(首页)", "even": "(偶数页)"}
for el in _root(zf, "word/document.xml").iter():
if el.tag in (W + "headerReference", W + "footerReference"):
tgt = rels.get(el.get(R + "id"), "").lstrip("/")
path = tgt if tgt.startswith("word/") else "word/" + tgt
word = "页眉" if el.tag == W + "headerReference" else "页脚"
labels.setdefault(path, word + kinds.get(el.get(W + "type", "default"), ""))
return labels
def extract_docx(path: Path) -> List[Item]:
"""抽取 docx 全部文字,返回 [(位置标签, 文字)]。"""
items: List[Item] = []
with zipfile.ZipFile(path) as zf:
labels = _hf_labels(zf)
for name in zf.namelist():
if not re.fullmatch(r"word/[^/]+\.xml", name):
continue
base = name[5:-4]
if name == "word/document.xml":
items.extend(_paragraphs(_root(zf, name), "正文"))
elif re.fullmatch(r"(header|footer)\d*", base):
default = "页眉" if base.startswith("header") else "页脚"
items.extend(_paragraphs(_root(zf, name), labels.get(name, default)))
elif base in ("footnotes", "endnotes", "comments"):
tag = {"footnotes": "脚注", "endnotes": "尾注", "comments": "批注"}[base]
items.extend(_paragraphs(_root(zf, name), tag))
return items
def _stext(el: ET.Element) -> str:
return "".join(t.text or "" for t in el.iter(S + "t"))
def extract_xlsx(path: Path) -> List[Item]:
items: List[Item] = []
with zipfile.ZipFile(path) as zf:
for name in zf.namelist():
if not name.startswith("xl/") or not name.endswith(".xml"):
continue
if name == "xl/workbook.xml":
for sh in _root(zf, name).iter(S + "sheet"):
items.append(("工作表名", sh.get("name", "")))
elif name == "xl/sharedStrings.xml":
for si in _root(zf, name).iter(S + "si"):
items.append(("单元格", _stext(si)))
elif re.fullmatch(r"xl/comments\d*\.xml", name):
for c in _root(zf, name).iter(S + "comment"):
items.append(("批注", _stext(c)))
elif re.fullmatch(r"xl/worksheets/sheet\d+\.xml", name):
sheet = name.rsplit("/", 1)[1][:-4]
for el in _root(zf, name).iter():
tag = el.tag.replace(S, "")
if tag in ("oddHeader", "evenHeader", "firstHeader"):
items.append(("页眉(%s)" % sheet, el.text or ""))
elif tag in ("oddFooter", "evenFooter", "firstFooter"):
items.append(("页脚(%s)" % sheet, el.text or ""))
elif tag == "is": # 内联字符串
items.append(("单元格(%s)" % sheet, _stext(el)))
return items
def extract(path: Path) -> List[Item]:
suffix = path.suffix.lower()
if suffix == ".docx":
return extract_docx(path)
if suffix == ".xlsx":
return extract_xlsx(path)
data = path.read_bytes()
try:
text = data.decode("utf-8-sig")
except UnicodeDecodeError:
text = data.decode("gb18030", errors="replace")
return [("第%d行" % i, line) for i, line in enumerate(text.splitlines(), 1) if line.strip()]
def scan(items: List[Item]) -> List[Tuple[str, str, str]]:
"""返回 [(位置, 类型, 命中词)]。"""
hits = []
for label, text in items:
flat = re.sub(r"\s+", " ", text)
for kind, rx in PATTERNS:
for m in rx.finditer(flat):
hits.append((label, kind, m.group(0)))
return hits
def main(argv=None) -> int:
for stream in (sys.stdout, sys.stderr):
try:
stream.reconfigure(encoding="utf-8")
except Exception:
pass
ap = argparse.ArgumentParser(description="检查文件中的保密标识(含页眉页脚、文本框、表格),命中退出码 1。")
ap.add_argument("files", nargs="+", help="txt、md、docx、xlsx 文件")
args = ap.parse_args(argv)
code = 0
for f in args.files:
path = Path(f)
try:
items = extract(path)
except (OSError, zipfile.BadZipFile, ET.ParseError, ValueError, KeyError) as exc:
print("[读取失败] %s:%s" % (path.name, exc))
code = 2
continue
hits = scan(items)
if not hits:
print("[未发现保密标识] " + path.name)
continue
if code == 0:
code = 1
print("[发现保密标识] %s:共 %d 处,请立即停止处理并按保密红线告知用户" % (path.name, len(hits)))
seen = set()
for label, kind, word in hits:
if (label, word) not in seen:
seen.add((label, word))
print(" [%s] %s:%s" % (label, kind, word))
return code
if __name__ == "__main__":
sys.exit(main())
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |