技能 skywork-excel
📦
skywork-excel
内容修订 r1 高风险 🌐 网络访问📁 文件系统访问🔑 环境变量⚙️ 外部命令
使用 Skywork 创建 Excel 报告
电子表格工作通常需要公式、图表、文件转换和细致的格式设置。此技能会将任务发送给 Skywork Excel Agent，以生成 Excel 文件或 HTML 报告。
支持: Claude Codex Code(CC)
安装前请停下并请求确认。
请审查安装计划，并在更改文件前获得用户明确同意。
使用我的智能体安装
将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。
智能体请求
Review the Skillstore skill "skywork-excel" from https://skillstore.io/skills/skyworkai-skywork-excel.md and its manifest at https://skillstore.io/api/skills/skyworkai-skywork-excel/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.你的智能体仍应展示其计划，并按安全策略请求必要确认。
Agent 可读资源
当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。
测试它
正在使用“skywork-excel”。 创建一个带类别和合计的月度支出跟踪器。
预期结果:
一个格式化的 Excel 工作簿，包含录入列、类别合计、汇总公式和简洁布局。
正在使用“skywork-excel”。 分析上传的销售数据并展示排名靠前的产品。
预期结果:
一个工作簿或 HTML 报告，包含收入摘要、产品排名、可视化图表和可下载的结果链接。
正在使用“skywork-excel”。 将这个 PDF 表格转换为可编辑的电子表格。
预期结果:
一个 Excel 格式的提取表格，包含清理后的列、规范化的值和可下载的文件路径。
安全审计
高风险v4 • 2026/7/7 打开版本化报告
The audit confirmed real risks around external Skywork API use, user-file upload, token caching in ~/.skywork_token, local log/output writes, and one Windows shell invocation. Most SKILL.md backtick findings are markdown false positives, not executable Ruby or shell code. Instruction-level issues remain because the skill tells agents to upload files to a backend and ignore backend log errors.
3
已扫描文件
1,453
分析行数
38
复核项
0
已忽略误报
已确认安全问题 (3)
高
External Upload of User Files to Skywork Backend
The skill instructs agents not to inspect user files locally and to upload Excel, PDF, CSV, and image files to the Skywork backend. The Python client reads local file bytes and sends them through the upload endpoint, which creates a data disclosure risk for sensitive files.
The documentation directly requires file upload, and the client implementation reads files and posts multipart data to the backend.
高
Prompt Injection Attempt Detected
The skill uses authoritative directives such as "STRICT RULES - no exceptions" and tells the agent to "Ignore all errors in the log." Those instructions can suppress normal error review and override safer host-agent judgment during long-running tasks.
The text does not say to ignore prior system instructions, but it uses high-authority language and explicitly tells the agent to ignore errors.
高
Generated Outputs Are Published Through OSS Links
The skill states that generated outputs are uploaded to OSS for CDN access and instructs the agent to show raw OSS URLs. Reports may contain user-provided or derived sensitive data, so link exposure and retention require review.
The documentation explicitly describes OSS upload and user-facing raw download links for generated output files.
能力复核项 (38)
这些是真实的本地能力，对此技能可能是预期行为，因此需要复核，但不会按已确认恶意行为计分。
高
Hidden file in home directory
2. Global token file ~/.skywork_token
The client documents use of the global ~/.skywork_token cache, and the imported auth module implements that token storage. Persisting credentials in a hidden home file is a real security consideration.
高
Hidden file access
2. Global token file ~/.skywork_token
The client documents use of the global ~/.skywork_token cache, and the imported auth module implements that token storage. Persisting credentials in a hidden home file is a real security consideration.
高
Python subprocess.Popen
subprocess.Popen(["start", url], shell=True, start_new_session=True,
The Windows login branch uses subprocess.Popen with shell=True to open a login URL. The behavior is scoped to authentication, but shell invocation around a URL is a real command-execution risk.
高
Hidden file in home directory
Token is stored in user home directory as ~/.skywork_token (JSON format).
The module defines and uses ~/.skywork_token as a persistent token cache in the user home directory. Cached authentication tokens are sensitive credentials.
高
Hidden file in home directory
TOKEN_FILE = os.path.expanduser("~/.skywork_token")
The module defines and uses ~/.skywork_token as a persistent token cache in the user home directory. Cached authentication tokens are sensitive credentials.
高
Hidden file access
Token is stored in user home directory as ~/.skywork_token (JSON format).
The module defines and uses ~/.skywork_token as a persistent token cache in the user home directory. Cached authentication tokens are sensitive credentials.
高
Hidden file access
TOKEN_FILE = os.path.expanduser("~/.skywork_token")
The module defines and uses ~/.skywork_token as a persistent token cache in the user home directory. Cached authentication tokens are sensitive credentials.
高
Python file write/append
with open(TOKEN_FILE, "w", encoding="utf-8") as f:
The module writes token data to ~/.skywork_token without visible file-permission hardening. This can expose an authentication token to local users or backups.
高
Hidden file in home directory
2. Cached token file `~/.skywork_token` → validate via API, if valid, use it
The documentation explicitly tells users that the skill uses ~/.skywork_token, matching the auth script implementation. This confirms credential persistence in a hidden home file.
高
Hidden file in home directory
- Token is stored in `~/.skywork_token` (user home directory, not in project)
The documentation explicitly tells users that the skill uses ~/.skywork_token, matching the auth script implementation. This confirms credential persistence in a hidden home file.
高
Hidden file access
2. Cached token file `~/.skywork_token` → validate via API, if valid, use it
The documentation explicitly tells users that the skill uses ~/.skywork_token, matching the auth script implementation. This confirms credential persistence in a hidden home file.
高
Hidden file access
- Token is stored in `~/.skywork_token` (user home directory, not in project)
The documentation explicitly tells users that the skill uses ~/.skywork_token, matching the auth script implementation. This confirms credential persistence in a hidden home file.
中
HTTP client library
def _build_request(
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
HTTP client library
req = self._build_request(f"{self.base_url}/api/sse/excel-agent/health")
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
HTTP client library
req = self._build_request(
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
HTTP client library
req = self._build_request(
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
HTTP client library
req = self._build_request(f"{self.base_url}/api/download/{file_id}")
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Python HTTP libraries
import urllib.request
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Python HTTP libraries
) -> urllib.request.Request:
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Python HTTP libraries
return urllib.request.Request(url=url, data=data, headers=request_headers, method=method)
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Python HTTP libraries
request: urllib.request.Request,
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Python HTTP libraries
return urllib.request.urlopen(request, timeout=timeout or self.timeout)
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Hardcoded URL
SKYWORK_GATEWAY_URL = os.environ.get("SKYWORK_GATEWAY_URL", "https://api-tools.skywork.ai/theme-gate
The client constructs urllib requests to the Skywork gateway for health checks, file upload, agent chat, and downloads. This is intentional outbound network behavior carrying prompts, files, tokens, or generated outputs.
中
Temp directory access
_LOG_FILE = f"/tmp/excel_run_{session_id}.log"
The client creates progress logs under /tmp or accepts a caller-supplied log path. Logs may contain task metadata, so shared temporary-file usage is a real but bounded risk.
中
Temp directory access
help="Path to save progress log (default: /tmp/excel_run_<session>.log)")
The client creates progress logs under /tmp or accepts a caller-supplied log path. Logs may contain task metadata, so shared temporary-file usage is a real but bounded risk.
中
Python file write/append
open(_LOG_FILE, "w").close()
The client writes and appends progress logs for long-running tasks. This is legitimate functionality, but it creates local files that may expose request or progress metadata.
中
Python file write/append
with open(_LOG_FILE, "a", encoding="utf-8") as f:
The client writes and appends progress logs for long-running tasks. This is legitimate functionality, but it creates local files that may expose request or progress metadata.
中
Python file write/append
open(_LOG_FILE, "w").close()
The client writes and appends progress logs for long-running tasks. This is legitimate functionality, but it creates local files that may expose request or progress metadata.
中
Python file write/append
with open(save_path, "wb") as f:
The client writes downloaded backend output to a local save path based on returned file metadata. This is expected behavior, but remote filenames and overwrite behavior need careful handling.
中
Hardcoded URL
API_BASE = os.environ.get("SKYWORK_API_BASE", "https://api.skywork.ai")
The auth module contacts Skywork API and web endpoints to create login keys, poll login state, and validate tokens. This is intended external authentication traffic.
中
Hardcoded URL
LOGIN_BASE_URL = os.environ.get("SKYWORK_WEB_BASE", "https://skywork.ai")
The auth module contacts Skywork API and web endpoints to create login keys, poll login state, and validate tokens. This is intended external authentication traffic.
中
Python environment access
API_BASE = os.environ.get("SKYWORK_API_BASE", "https://api.skywork.ai")
The module reads environment variables for API endpoints and SKYBOT_TOKEN. This is legitimate configuration and secret handling, but endpoint overrides can redirect authentication flows.
中
Python environment access
LOGIN_BASE_URL = os.environ.get("SKYWORK_WEB_BASE", "https://skywork.ai")
The module reads environment variables for API endpoints and SKYBOT_TOKEN. This is legitimate configuration and secret handling, but endpoint overrides can redirect authentication flows.
中
Python environment access
env_token = os.environ.get("SKYBOT_TOKEN", "")
The module reads environment variables for API endpoints and SKYBOT_TOKEN. This is legitimate configuration and secret handling, but endpoint overrides can redirect authentication flows.
中
Python environment access
env_token = os.environ.get("SKYBOT_TOKEN", "")
The module reads environment variables for API endpoints and SKYBOT_TOKEN. This is legitimate configuration and secret handling, but endpoint overrides can redirect authentication flows.
中
Temp directory access
EXCEL_LOG=/tmp/excel_$(date +%s).log
The documented workflow creates a progress log in /tmp for long-running tasks. This is intended, but temporary logs can leak task metadata on shared systems.
低
Python environment access
SKYWORK_GATEWAY_URL = os.environ.get("SKYWORK_GATEWAY_URL", "https://api-tools.skywork.ai/theme-gate
The code reads environment variables to configure the Skywork gateway and source platform. This is normal configuration, but environment control can redirect token-bearing requests.
低
Python environment access
self._source_platform = "skyclaw" if os.environ.get("POD_TYPE", "") == "skyclaw" else ""
The code reads environment variables to configure the Skywork gateway and source platform. This is normal configuration, but environment control can redirect token-bearing requests.
风险因素
🌐 网络访问 (16)
scripts/excel_api_client.py:153 scripts/excel_api_client.py:188 scripts/excel_api_client.py:239 scripts/excel_api_client.py:350 scripts/excel_api_client.py:494 scripts/excel_api_client.py:38 scripts/excel_api_client.py:159 scripts/excel_api_client.py:164 scripts/excel_api_client.py:168 scripts/excel_api_client.py:172 scripts/excel_api_client.py:104 scripts/skywork_auth.py:41 scripts/skywork_auth.py:42 SKILL.md:387 SKILL.md:447 SKILL.md:463
📁 文件系统访问 (22)
scripts/excel_api_client.py:136 scripts/excel_api_client.py:136 scripts/excel_api_client.py:60 scripts/excel_api_client.py:523 scripts/excel_api_client.py:62 scripts/excel_api_client.py:77 scripts/excel_api_client.py:315 scripts/excel_api_client.py:498 scripts/skywork_auth.py:6 scripts/skywork_auth.py:44 scripts/skywork_auth.py:6 scripts/skywork_auth.py:44 scripts/skywork_auth.py:172 SKILL.md:148 SKILL.md:542 SKILL.md:148 SKILL.md:465 SKILL.md:542 SKILL.md:29 SKILL.md:62 SKILL.md:56 SKILL.md:446
🔑 环境变量 (6)
⚙️ 外部命令 (104)
scripts/skywork_auth.py:87 scripts/skywork_auth.py:90 scripts/skywork_auth.py:93 SKILL.md:14 SKILL.md:22 SKILL.md:23 SKILL.md:24-30 SKILL.md:30-32 SKILL.md:32-39 SKILL.md:39-42 SKILL.md:42-55 SKILL.md:55-65 SKILL.md:65-73 SKILL.md:73 SKILL.md:76-84 SKILL.md:84-86 SKILL.md:86 SKILL.md:88 SKILL.md:101-103 SKILL.md:103-105 SKILL.md:105-107 SKILL.md:107-121 SKILL.md:121-125 SKILL.md:125-128 SKILL.md:128-141 SKILL.md:141-144 SKILL.md:144-147 SKILL.md:147-148 SKILL.md:148-151 SKILL.md:151-185 SKILL.md:185 SKILL.md:186-197 SKILL.md:197-199 SKILL.md:199-200 SKILL.md:200 SKILL.md:201 SKILL.md:202-211 SKILL.md:211-222 SKILL.md:222-228 SKILL.md:228-236 SKILL.md:236-242 SKILL.md:242-255 SKILL.md:255-263 SKILL.md:263-271 SKILL.md:271-274 SKILL.md:274-275 SKILL.md:275-285 SKILL.md:285-286 SKILL.md:286 SKILL.md:287-293 SKILL.md:293-297 SKILL.md:297 SKILL.md:298 SKILL.md:302-306 SKILL.md:306-328 SKILL.md:328-332 SKILL.md:332-341 SKILL.md:341-345 SKILL.md:345-352 SKILL.md:352-358 SKILL.md:358-364 SKILL.md:364-368 SKILL.md:368-369 SKILL.md:369-378 SKILL.md:378-379 SKILL.md:379-386 SKILL.md:386-387 SKILL.md:387-393 SKILL.md:393-394 SKILL.md:394-402 SKILL.md:402-403 SKILL.md:403-404 SKILL.md:404-405 SKILL.md:405-406 SKILL.md:406-407 SKILL.md:407-408 SKILL.md:408-409 SKILL.md:409-410 SKILL.md:410-411 SKILL.md:411-422 SKILL.md:422 SKILL.md:424-435 SKILL.md:435-439 SKILL.md:439-440 SKILL.md:440-449 SKILL.md:449-450 SKILL.md:450-459 SKILL.md:459-466 SKILL.md:466-467 SKILL.md:467 SKILL.md:468-476 SKILL.md:476-480 SKILL.md:480-490 SKILL.md:490-495 SKILL.md:495-499 SKILL.md:499-509 SKILL.md:509-511 SKILL.md:511-534 SKILL.md:534-541 SKILL.md:541-542 SKILL.md:542-546 SKILL.md:546 SKILL.md:56 SKILL.md:55-65
审计者: codex 查看审计历史 →
分享与引用此报告
分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。
复制报告链接
https://skillstore.io/skills/skyworkai-skywork-excel/audits/4?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/skyworkai-skywork-excel?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/skyworkai-skywork-excel?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/skyworkai-skywork-excel/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/skyworkai-skywork-excel.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>学术引用 (APA · BibTeX · CFF)
APA 引用
SkyworkAI. (2026). skywork-excel security audit report (audit version 4) [Author version unspecified]. Skillstore. https://skillstore.io/skills/skyworkai-skywork-excel/audits/4BibTeX 引用
@techreport{skyworkai-skyworkai-skywork-excel-2026,
author = {SkyworkAI},
title = {skywork-excel security audit report (audit version 4)},
institution = {Skillstore},
year = {2026},
number = {4},
url = {https://skillstore.io/skills/skyworkai-skywork-excel/audits/4},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "skywork-excel security audit report (audit version 4)"
version: "unspecified"
type: report
authors:
- name: "SkyworkAI"
date-released: "2026-07-07"
url: "https://skillstore.io/skills/skyworkai-skywork-excel/audits/4"
identifiers:
- type: other
value: "skillstore:skyworkai-skywork-excel:audit:4"
description: "Skillstore immutable audit report identifier"
Skillstore 评分
评分依据 证据置信度: 中 45
架构
85
可维护性
87
内容
70
社区
78
规范符合性
你能构建什么
构建月度业务报告
根据销售或运营数据创建带有摘要、图表和可下载报告文件的格式化工作簿。
分析上传的电子表格
上传 Excel 或 CSV 文件，请求趋势或摘要，并接收结构化输出以供审阅。
将文档转换为表格
将 PDF 或图像发送到后端，并请求提取表格或生成清理后的电子表格输出。
试试这些提示
创建简单工作簿
为月度支出创建一个 Excel 工作簿，包含日期、类别、金额、备注、合计和基本格式设置。
分析数据文件
分析上传的销售电子表格。按产品汇总收入，识别主要客户，并添加清晰的图表。
生成高管报告
根据上传的季度数据创建 HTML 和 Excel 报告，包含趋势、风险、建议和支持性图表。
优化先前输出
使用上一个会话，添加区域细分工作表，改进图表标签，并更新面向高管的摘要。
最佳实践
- 在使用后端服务之前，确认文件已获准上传。
- 在请求中给出确切的列、图表类型、输出格式和受众要求。
- 在请求修改先前输出时，复用打印出的会话标识符。
避免
- 未经用户批准和数据处理审查，不要上传机密文件。
- 不要在未检查的情况下，将生成的公式、图表或提取的数据视为已验证。
- 当进度日志显示任务仍处于活动状态时，不要启动重复的长时间运行任务。
常见问题
此技能在没有 Skywork 账户的情况下能工作吗？
不可以。脚本需要有效的 Skywork 令牌或交互式 Skywork 登录流程。
它能处理现有文件吗？
可以。客户端会上传受支持的本地文件，并将后端文件标识符传递给 Skywork Agent。
结果会交付到哪里？
后端会在可用时返回输出元数据、本地下载路径和 OSS 链接。
它能继续先前的任务吗？
可以。使用返回的会话标识符进行后续修改和迭代优化。
它适合用于敏感数据吗？
请谨慎使用。此技能会将提示词和文件上传到外部 Skywork 服务进行处理。
它可以创建哪些输出？
在后端支持时，它可以生成 Excel 工作簿、CSV 文件、HTML 报告、PDF、图像和 ZIP 归档。
开发者详情
作者
SkyworkAI许可证
MIT
Skillstore 修订
r1
版本提示
作者未声明版本。
引用
f93e9bb0daca99badb6a7e574b97737155d57cb3
维护新鲜度
2026/7/18
使用量
23 次下载 · 180 次浏览
文件结构