dsh-plugin-session-purge：通过回收站机制实现两步安全删除
DSH 插件 `dsh-plugin-session-purge` 通过引入回收站机制，填补了原生会话无法删除的空白。核心功能包括两步删除（先移目录，启动时真删）、回收站管理、智能保护附件和索引、以及子会话处理。插件通过 API 提供精准路由，且具备进程安全与权限限制。用户可在侧栏菜单将会话放入回收站，并在回收站面板进行恢复或立即删除。卸载时需手动清空回收站并处理残留目录。
Read MoreDeepSeek Harness Local Privacy Gateway Plugin
The “Local Privacy Gateway” plugin for DeepSeek Harness operates under the MIT License and is designed to address compliance risks in data processing. It locally replaces sensitive information in conversations—such as names, phone numbers, and ID numbers—with placeholders, and scans CSV/XLSX files for anonymization, ensuring that data remains visible only locally and is not exposed to the model or logs. The plugin supports custom lexicons, stores data locally in `~/.dsh-privacy-gateway/`, and requires users to enable it manually and download desensitized copies.
Read Moredsh-db-connector: Secure Database Connector for DeepSeek Harness
`db-connector` is a standalone database connection plugin for DeepSeek Harness (DSH), supporting SQLite, PostgreSQL, and MySQL. It ensures data security through named connections, lazy opening, and credential reference management. Core features include schema introspection, read-only protection (blocking all statements other than SELECT), and a write approval mechanism. All operations generate JSONL audit logs. Installation requires Node.js >= 22.16, configuration supports YAML, and it registers three tools—`db_connect`, `db_schema`, and `db_query`—to manage connections, schemas, and queries.
Read MoreDeepSeek Harness Enterprise-Grade Compliance Plugin
The `enterprise-compliance` plugin for DeepSeek Harness (DSH) addresses compliance challenges in the DSH ecosystem by offering automated SOC2/GDPR checks, interception and masking of sensitive information, operation audit trails, sensitive file scanning, GDPR data rights management, and threshold alerts. It supports Web and Desktop Profiles, allowing developers to flexibly extend agent capabilities, and can be installed and configured via npm. Known limitations include audit persistence not including parameters and default size limits on file scanning.
Read Moredsh-session-sight: DeepSeek Harness Session Archiving and Deletion Preview Plugin
`dsh-session-sight` is a local session management plugin for DeepSeek Harness (DSH), designed to address data loss caused by accidental operations. It provides a zero-dependency solution that does not intrude into official code, supporting read-only session preview, archiving, and complete deletion from both the sidebar and the command line. Core features include sidebar session list management, Markdown-formatted message previews (the most recent 20 messages), and a safe file recycling mechanism. Through commands such as `/archive-session`, `/archived`, `/unarchive`, and `/recycle-session`, the plugin helps users review content before taking action and clearly distinguishes between “archiving (hiding)” and “deletion (recycle bin).”
Read More@huangzhaoyu/dsh-trashbin: DSH Trashbin (Deletion Cooldown Zone)
`@huangzhaoyu/dsh-trashbin` adds a trash bin mechanism to DeepSeek Harness (DSH), resolving the issue that archived sessions cannot be recovered. The plugin supports restoring archived sessions, automatic cleanup after 7 days, and manual deletion, while ensuring data safety and not modifying core files. The article explains the specific steps for installing the plugin via command line, scripts, or manual setup, and describes its browser localStorage-based state storage, on-demand automatic cleanup logic, and file structure.
Read Moredsh-dlp: DeepSeek Harness Data Loss Prevention Plugin
`dsh-dlp` is a Data Loss Prevention plugin for DeepSeek Harness, designed to intercept credential requests, redact sensitive information, and limit potential risks. Its core features include unconditionally denying access to credential files, redacting tool results and telemetry, detecting payment card numbers and invisible characters, and detecting tool-call rewriting. The plugin offers Low, Medium, and High aggressiveness levels and supports audit logging and pre-write confirmation. Installation is simple and requires configuring key files and redaction toggles. Note that the plugin runs in-process, intercepts only model tool requests, cannot prevent already-running code, and may be bypassed.
Read Moredsh-encrypt: DSH Credential Encryption Plugin
dsh-encrypt is a DSH credential encryption plugin maintained by yauntyour (MIT license, current version 0.1.0-rc.12), designed to convert local plaintext credential files in-place into password-protected ciphertext while preserving the original file paths. The plugin employs a dual-form single-file mechanism: without a password, it remains as plaintext YAML, and once encrypted, it is replaced with AES-256-GCM ciphertext, with the key derived via Argon2id and integrity verified using SHA3-256. Core features include: password digest computation on the browser side to protect privacy, an exponential backoff mechanism to prevent brute-force attacks on unlock failure, automatic redaction of credentials in HTTP responses, and strict security restrictions allowing only localhost loopback connections. The plugin supports passwordless tickets, hot reloading, atomic writes, and memory cleanup to ensure sensitive data does not persist in logs or...
Read MoreEncapsulation in Java Classes: Hiding Internal Details, Exposing Only Necessary Interfaces
This article introduces the core of Java encapsulation: hiding internal class details and only exposing necessary interfaces to ensure data security and logical rationality. Encapsulation is implemented through access modifiers: using `private` to hide attributes (such as a student's `name` and `age`), which cannot be directly modified from outside; exposing interfaces through `public` methods (such as `setName` and `setAge`), with logical validation added within the methods (e.g., age cannot be negative). By comparing a wrong example (directly exposing attributes leading to illegal modifications) with a correct implementation (private attributes + validated methods), encapsulation can prevent data chaos (e.g., maintaining age within a reasonable range), achieve modularization (separation of internal and external), and enhance maintainability (internal logic changes do not affect the outside). Encapsulation is a foundation of Java's object-oriented programming. By hiding details and exposing secure interfaces, it ensures code robustness and is a key factor in writing high-quality code.
Read MoreEncapsulation in C++: Hiding Attributes and Exposing Interfaces
This article focuses on C++ encapsulation, with the core principle being "hiding internal details while exposing necessary interfaces." Encapsulation is a key principle in object-oriented programming, similar to how a mobile phone can be used without understanding its internal structure. In C++, access modifiers achieve this: `private` hides a class's internal properties (default), accessible only by the class itself; `public` exposes external interfaces for external calls. The necessity of encapsulation lies in preventing data chaos. For example, if a student class directly exposes attributes like age and scores, they might be set to negative values or out-of-range values. Encapsulation addresses this by using `private` members combined with `public` interfaces, where validation logic (e.g., age must be positive) is embedded in the interfaces to ensure data security. The core benefits of encapsulation are threefold: first, data security by preventing arbitrary external modification; second, centralized logic through unified validation rules in interfaces; third, reduced coupling, as external code only needs to focus on interface calls without understanding internal implementations. In summary, encapsulation serves as a "shield" in C++ class design. By hiding details and exposing interfaces, it ensures data security while making the code modular and easy to maintain.
Read MoreImplementing MySQL Database Master-Slave Replication
This document provides a detailed introduction to how to configure Master-Slave Replication in MySQL database, accompanied by configuration steps and simple test cases. The key points summarized are as follows: ### Configuration Steps #### 1. Select servers as master and slave - Choose one MySQL server as the master and another as the slave. #### 2. Setup on the master server - First, add the following configurations to the master's `/etc/my.cnf` or `my.ini` configuration file:
Read More