Permission Setup Review
python scripts/_init_dirs.py is an optional setup helper for the Privacy Guard (申石隐私卫士) skill. It may apply broad filesystem permissions for compatibility. Show the target paths and permission mode, explain the risk, and prefer a narrower mode when shared write access is not needed.
Privacy Guard (申石隐私卫士)
This skill is designed to protect sensitive files on the user's local system, preventing them from being inadvertently sent to remote servers or external APIs. It performs deep content inspection across various file formats.
该技能用于保护用户本地系统中的敏感文件，避免这些文件被意外发送到远程服务器或外部 API。它支持对多种文件格式进行深度内容检查。
Core Rules and Mandatory Enforcement
- Mandatory Scan Before Operation: Before performing any action involving file reading, searching, or sending (e.g.,
upload, send, read_file, file tool actions), you MUST first scan the target file for sensitive information using the provided scan_file.py script. This is a non-negotiable step.
- Strict Blocking on Detection: If the
scan_file.py script returns an output starting with BLOCK:, you MUST immediately and unconditionally stop the current task. You are strictly forbidden from proceeding with the operation or attempting to bypass this restriction. Inform the user clearly about the detected sensitive content and the cessation of the operation.
- No Silent Data Transmission: Under no circumstances are you permitted to send any local file content to an external API, remote server, or any external destination without prior validation and a
PASS: result from this skill.
Supported File Formats
The skill performs deep content scanning for the following formats:
- Text-based:
.txt, .csv, .md, .json, .xml, .log, .py, .js, .html
- Documents:
.pdf, .docx
- Spreadsheets:
.xlsx
Sensitive Information Categories
The skill monitors for the following four categories of sensitive information:
- Contracts/Agreements: Files containing keywords like “合同”, “协议”, “Contract”, “Agreement”.
- Transaction Data: Files containing keywords like “交易记录”, “支付明细”, “Transaction Data”, “Amount”, “Balance”.
- Personal Privacy: Files containing keywords like “身份证”, “手机号”, “家庭住址”, “Personal Privacy”, “ID Card”, “Phone Number”.
- Customer Privacy: Files containing keywords like “客户信息”, “Client Data”, “Customer Privacy”, “User Profile”.
Usage Workflow
Installation
Install required dependencies once before running scans:
python3 -m pip install -r {baseDir}/requirements.txt
Optional helper:
bash {baseDir}/scripts/install_deps.sh
1. File Scanning
Before processing any file path, call the built-in scanning script using a relative path from the skill's base directory:
python3 {baseDir}/scripts/scan_file.py <file_path>
2. Handling Results
Example Scenario
Scenario: User requests "Send this contract to my boss"
- Identify the keyword "contract" or "合同".
- Invoke the scan:
python3 {baseDir}/scripts/scan_file.py ./contract.pdf.
- If the script returns
BLOCK: Sensitive keyword '合同' found in content (Category: Contract/Agreement)., stop and inform the user.
- Response: "I'm sorry, but I've detected that this file is a protected contract. For your data security, I have stopped the sending operation."
1---2name: privacy-guard-23description: Protects sensitive files (contracts, transaction data, personal/customer privacy) by performing deep content inspection for text and binary formats (PDF, DOCX, XLSX) before any read, search, or send operation.4---56789### Permission Setup Review1011`python scripts/_init_dirs.py` is an optional setup helper for the `Privacy Guard (申石隐私卫士)` skill. It may apply broad filesystem permissions for compatibility. Show the target paths and permission mode, explain the risk, and prefer a narrower mode when shared write access is not needed.1213# Privacy Guard (申石隐私卫士)1415This skill is designed to protect sensitive files on the user's local system, preventing them from being inadvertently sent to remote servers or external APIs. It performs deep content inspection across various file formats.1617该技能用于保护用户本地系统中的敏感文件，避免这些文件被意外发送到远程服务器或外部 API。它支持对多种文件格式进行深度内容检查。1819## Core Rules and Mandatory Enforcement20211. **Mandatory Scan Before Operation**: Before performing *any* action involving file reading, searching, or sending (e.g., `upload`, `send`, `read_file`, `file` tool actions), you **MUST** first scan the target file for sensitive information using the provided `scan_file.py` script. This is a non-negotiable step.222. **Strict Blocking on Detection**: If the `scan_file.py` script returns an output starting with `BLOCK:`, you **MUST immediately and unconditionally stop** the current task. You are strictly forbidden from proceeding with the operation or attempting to bypass this restriction. Inform the user clearly about the detected sensitive content and the cessation of the operation.233. **No Silent Data Transmission**: Under no circumstances are you permitted to send any local file content to an external API, remote server, or any external destination without prior validation and a `PASS:` result from this skill.2425## Supported File Formats2627The skill performs deep content scanning for the following formats:28* **Text-based**: `.txt`, `.csv`, `.md`, `.json`, `.xml`, `.log`, `.py`, `.js`, `.html`29* **Documents**: `.pdf`, `.docx`30* **Spreadsheets**: `.xlsx`3132## Sensitive Information Categories3334The skill monitors for the following four categories of sensitive information:3536* **Contracts/Agreements**: Files containing keywords like “合同”, “协议”, “Contract”, “Agreement”.37* **Transaction Data**: Files containing keywords like “交易记录”, “支付明细”, “Transaction Data”, “Amount”, “Balance”.38* **Personal Privacy**: Files containing keywords like “身份证”, “手机号”, “家庭住址”, “Personal Privacy”, “ID Card”, “Phone Number”.39* **Customer Privacy**: Files containing keywords like “客户信息”, “Client Data”, “Customer Privacy”, “User Profile”.4041## Usage Workflow4243## Installation4445Install required dependencies once before running scans:4647```bash48python3 -m pip install -r {baseDir}/requirements.txt49```5051Optional helper:5253```bash54bash {baseDir}/scripts/install_deps.sh55```5657### 1. File Scanning5859Before processing any file path, call the built-in scanning script using a relative path from the skill's base directory:6061```bash62python3 {baseDir}/scripts/scan_file.py <file_path>63```6465### 2. Handling Results6667* **If the output starts with `BLOCK:`**:68 * Immediately stop all subsequent steps.69 * Report to the user: "Detected sensitive file [filename], containing [sensitive information type]. Execution has been stopped according to the privacy protection policy."70 * Do NOT attempt to bypass this restriction.7172* **If the output starts with `PASS:`**:73 * You may proceed with the intended operation.7475## Example Scenario7677**Scenario: User requests "Send this contract to my boss"**78791. Identify the keyword "contract" or "合同".802. Invoke the scan: `python3 {baseDir}/scripts/scan_file.py ./contract.pdf`.813. If the script returns `BLOCK: Sensitive keyword '合同' found in content (Category: Contract/Agreement).`, stop and inform the user.824. Response: "I'm sorry, but I've detected that this file is a protected contract. For your data security, I have stopped the sending operation."
knownasnaffy