技能 python-executor
📦
python-executor
內容修訂 r1 高風險 ⚙️ 外部命令🌐 網路存取
在遠端沙盒中執行 Python 任務
透過 AI 助理執行 Python 可能需要設定、程式庫，以及安全的輸出處理。此技能會將 Claude、Codex 和 Claude Code 提示路由到 inference.sh Python 執行器，並搭配常見的資料、媒體和網頁程式庫。
支援: Claude Codex Code(CC)
安裝前請停止並請求確認。
請審查安裝計畫，並在變更檔案前取得使用者明確同意。
使用我的智慧代理安裝
將此請求複製給你的智慧代理，其中包含權威 Skill 頁面和清單。
智能體請求
Review the Skillstore skill "python-executor" from https://skillstore.io/skills/inference-skills-python-executor.md and its manifest at https://skillstore.io/api/skills/inference-skills-python-executor/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.你的智慧代理仍應顯示其計畫，並依安全政策請求必要確認。
Agent 可讀資源
當 AI Agent、爬蟲或腳本需要乾淨脈絡、而不是讀取完整頁面時，請使用這些連結。
測試它
正在使用「python-executor」。 分析小型銷售表並建立圖表。
預期結果:
- 總銷售額和平均銷售額的簡短文字摘要。
- 儲存在 outputs 目錄中的圖表檔案。
- 列出主要使用程式庫的備註。
正在使用「python-executor」。 建立簡單的漸層影像。
預期結果:
- 從 outputs 目錄回傳的已產生影像檔案。
- 確認影像尺寸和格式的精簡訊息。
正在使用「python-executor」。 從已核准的 API 擷取公開個人資料。
預期結果:
- 所選欄位的可讀摘要。
- 提醒已向核准端點發出對外請求。
安全審計
高風險v4 • 2026/7/5 開啟版本化報告
Most Markdown backtick findings are false positives from code fences, package names, or inline examples. The confirmed risks are documented belt commands that submit arbitrary Python to a remote executor, outbound HTTP examples, and related skill installation commands. No prompt injection text was found, but the safety claim is not verifiable from this skill file.
1
已掃描檔案
186
分析行數
13
複核項目
0
已忽略的誤報
已確認的安全問題 (2)
高
Remote Arbitrary Python Execution Capability
The skill authorizes belt CLI use and instructs users to submit Python code to infsh/python-executor. This can execute broad automation and network-capable scripts through an external service.
The allowed-tools entry permits belt commands, and examples send arbitrary code to a named executor. The use case list explicitly says it can run any Python script.
中
Unverified Sandbox Safety Claim
The documentation says execution is safe and isolated, but the skill file contains no implementation or policy details to verify the sandbox boundary. Users must trust the remote service and CLI behavior.
The safety claim is explicit, while the repository shard only provides documentation. No implementation evidence is present in the scanned file.
能力複核項目 (13)
這些是真實的本機能力，對此技能而言可能是預期行為，因此需要複核，但不會按已確認的惡意行為計分。
中
Ruby/shell backtick execution
```bash
The location covers a bash quick-start block that runs belt login and belt app run against infsh/python-executor. It is a real external command path for submitting Python code to a remote executor.
中
Ruby/shell backtick execution
```bash
The bash example invokes belt app run and submits Python that fetches a website. This is real external command execution with outbound network behavior.
中
Ruby/shell backtick execution
```bash
The bash example invokes belt app run to execute a data analysis script remotely. The command is hardcoded, but it still executes user-supplied Python through an external CLI.
中
Ruby/shell backtick execution
```bash
The bash example invokes belt app run to execute image-processing code in the remote executor. This confirms external command use, though the sample input is benign.
中
Ruby/shell backtick execution
```bash
The bash example invokes belt app run with media-generation code and an extended timeout. This is legitimate documented functionality but still a real external command execution path.
中
Ruby/shell backtick execution
```bash
The bash example invokes belt app run to execute 3D processing code remotely. This is a confirmed external command capability rather than a Markdown-only artifact.
中
Ruby/shell backtick execution
```bash
The bash example invokes belt app run and submits Python that calls the GitHub API. It combines external command execution with outbound HTTP behavior.
中
Ruby/shell backtick execution
```bash
The variants section shows belt app run commands, including a high-memory remote executor variant. This is real documented external command usage.
中
Ruby/shell backtick execution
```bash
The related skills section shows npx skills add commands that install additional skills. Installation commands can alter the local tool environment and require user approval.
低
Python HTTP libraries
"code": "import requests\nfrom bs4 import BeautifulSoup\n\nresponse = requests.get(\"https://example
The example imports requests and fetches https://example.com inside code submitted to the executor. This confirms network-capable Python execution, though the target is benign.
低
Python HTTP libraries
"code": "import requests\nimport json\n\nresponse = requests.get(\"https://api.github.com/users/octo
The example imports requests and calls the GitHub API from executor code. This is an intended outbound HTTP capability and should be disclosed.
低
Hardcoded URL
"code": "import requests\nfrom bs4 import BeautifulSoup\n\nresponse = requests.get(\"https://example
The hardcoded URL is used inside a requests.get call in the scraping example. It demonstrates outbound network access from submitted Python code.
低
Hardcoded URL
"code": "import requests\nimport json\n\nresponse = requests.get(\"https://api.github.com/users/octo
The hardcoded GitHub API URL is used inside a requests.get call. It demonstrates outbound API access from the executor.
風險因素
⚙️ 外部命令 (40)
SKILL.md:15 SKILL.md:17-24 SKILL.md:24-31 SKILL.md:31-38 SKILL.md:38-45 SKILL.md:45-50 SKILL.md:50 SKILL.md:51 SKILL.md:52 SKILL.md:53-56 SKILL.md:56 SKILL.md:57 SKILL.md:60 SKILL.md:61 SKILL.md:64-65 SKILL.md:65 SKILL.md:66-69 SKILL.md:69 SKILL.md:70 SKILL.md:73 SKILL.md:74 SKILL.md:80-84 SKILL.md:84-88 SKILL.md:88-92 SKILL.md:92-96 SKILL.md:96-100 SKILL.md:100-104 SKILL.md:104-109 SKILL.md:109-113 SKILL.md:113-117 SKILL.md:117-121 SKILL.md:121-125 SKILL.md:125-129 SKILL.md:129-131 SKILL.md:131-137 SKILL.md:137-141 SKILL.md:141-147 SKILL.md:147-164 SKILL.md:164 SKILL.md:169-178
審計單位: codex 檢視審計紀錄 →
分享與引用此報告
分享版本化評估報告、中性徽章、嵌入卡片與引用。Skillstore 只報告證據，不判斷該 Skill 是否安全。
複製報告連結
https://skillstore.io/skills/inference-skills-python-executor/audits/4?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/inference-skills-python-executor?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/inference-skills-python-executor?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/inference-skills-python-executor/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/inference-skills-python-executor.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>學術引用 (APA · BibTeX · CFF)
APA 引用
inference-skills. (2026). python-executor security audit report (audit version 4) [Author version unspecified]. Skillstore. https://skillstore.io/skills/inference-skills-python-executor/audits/4BibTeX 引用
@techreport{inference-skills-inference-skills-python-executor-2026,
author = {inference-skills},
title = {python-executor security audit report (audit version 4)},
institution = {Skillstore},
year = {2026},
number = {4},
url = {https://skillstore.io/skills/inference-skills-python-executor/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: "python-executor security audit report (audit version 4)"
version: "unspecified"
type: report
authors:
- name: "inference-skills"
date-released: "2026-07-05"
url: "https://skillstore.io/skills/inference-skills-python-executor/audits/4"
identifiers:
- type: other
value: "skillstore:inference-skills-python-executor:audit:4"
description: "Skillstore immutable audit report identifier"
比較變體
10 個可安裝變體每位作者的技能都保留為獨立的可安裝項目。推薦變體依 Skillstore 證據排序。
為何此變體排名第一
Skillstore 使用量較高
作者 / 來源 Skillstore 評分 證據置信度 Skillstore 使用量 已更新
infsh-skills 推薦
infsh-skills-python-executor
Skillstore 評分 50
證據置信度 高
Skillstore 使用量 45
已更新
2026-09-09
inference-skills-python-executor
Skillstore 評分 38
證據置信度 中
Skillstore 使用量 43
已更新
2026-09-09
inference-sh-skills-python-executor
Skillstore 評分 38
證據置信度 中
Skillstore 使用量 43
已更新
2026-09-09
inference-sh-9-python-executor
Skillstore 評分 38
證據置信度 中
Skillstore 使用量 43
已更新
2026-09-09
halt-catch-fire-python-executor
Skillstore 評分 38
證據置信度 中
Skillstore 使用量 43
已更新
2026-09-09
Skillstore 評分
評分依據 證據置信度: 中 55
架構
85
可維護性
87
內容
68
社群
83
規範符合度
你可以打造什麼
分析小型資料集
處理 CSV 形式的資料、計算摘要指標，並在不需本機 Python 設定的情況下儲存圖表。
建立媒體素材
使用預先安裝的 Python 程式庫產生影像、短影片、PDF 或 3D 檔案。
建立自動化指令碼原型
執行短篇 Python 指令碼，用於 API 呼叫、檔案轉換和可重複的實用工具任務。
試試這些提示
執行簡單計算
Use python-executor to run a short Python calculation. Print the result and explain what ran.
摘要資料集
Use python-executor to analyze this table. Return summary statistics and save one chart in the outputs directory.
轉換影像
Use python-executor to resize and annotate the provided image. Save the processed image in the outputs directory.
建立多步驟自動化
Use python-executor to fetch approved API data, transform it with pandas, save a report, and summarize the output files.
最佳實務
- 執行前先審查程式碼，尤其是當它會讀取檔案或發出網路請求時。
- 將必要成品儲存在 outputs 目錄中，以便可靠回傳。
- 先使用較短的逾時時間，只有在較繁重的工作需要時才提高限制。
避免
- 除非外部執行器已獲准處理，否則不要提交密鑰、權杖或私人檔案。
- 不要執行不受信任的爬取程式碼，或從未知來源複製的命令。
- 不要將此技能用於 GPU 工作負載或長時間執行的服務。
常見問題
這會在本機執行 Python 嗎？
不會。此技能使用 inference.sh belt CLI 在外部執行器中執行程式碼。
哪些 AI 工具可以使用它？
報告列出 Claude、Codex 和 Claude Code 為支援的工具。
它可以發出網路請求嗎？
可以。範例包含以 requests 為基礎的網頁爬取和 API 呼叫。
產生的檔案應該儲存在哪裡？
請將檔案儲存在 outputs 目錄中，讓執行器可以回傳它們。
它支援 GPU 工作負載嗎？
不支援。文件記載的環境僅限 CPU。
使用前應該檢查什麼？
確認程式碼、輸入和網路端點都已獲准用於外部執行服務。
開發者詳細資訊
授權
MIT
Skillstore 修訂
r1
版本提示
作者未聲明版本。
引用
a25199bc7d6b82598536822d1738eb5d5f54025b
維護新鮮度
2026/7/18
使用量
3 次下載 · 63 次瀏覽
檔案結構
📄 SKILL.md