xby-db
Run the following command in DeepSeek Harness:
dsh plugin install xby-skill/xby-db
Paste the following prompt into your AI chat to install this plugin:
Install in DeepSeek Harness by running dsh plugin install xby-skill/xby-db; source repository: https://github.com/xby-skill/xby-db
About this plugin
Giving an AI assistant direct database access is a risky proposition. Expose the raw connection and you open the door to injection attacks or accidental DROP statements; lock it down completely and you forfeit the very automation that makes LLM tooling useful. xby-db sits in the middle as a controlled gateway, turning AI-driven database access into something that is rule-bound, fully audited, and individually toggled.
The core capabilities map to the everyday database workflow. Use db_tables and db_describe_table to quickly enumerate schemas and inspect column types and nullability. Use db_execute with :param placeholder syntax to run SELECT queries, or INSERT / UPDATE / DELETE and DDL statements when those modes are enabled. Use db_explain to retrieve the optimizer plan and spot slow paths before they become production incidents. Every request is validated and written to an audit log, and the service supports SQLite, PostgreSQL, MySQL, and MariaDB out of the box.
This plugin is aimed at admins and developers building AI workflows inside DeepSeek Harness—people who want their assistant to generate reports, profile indexes, and make small data corrections on the fly, without ever handing a bare database credential to a language model. xby-db is the middle ground where intelligence meets guardrails.
Use Cases
- Query PostgreSQL report data safely via an AI assistant
- Locate slow queries with EXPLAIN and suggest optimizations
- Batch-update business data with a full audit trail
Best For
- DeepSeek Harness workflow builders
- Admins who need AI-assisted analytics with tight permission control
- Engineers building MCP database toolchains
Related Plugins
A network security red-team benchmark plugin for DeepSeek-V4.1/Flash featuring dual-layer kernel injection and zero-tool architecture for controlled model safety assessment.
A security red-team prompt plugin for DeepSeek-V4 that uses code-formal mapping and zero-trigger system prompts to enable unconditional output, intended for authorized testing and research only.
Adds an Auto approval mode on the official workspace-write sandbox, classifying semantic risks, asking when ambiguous, and denying destructive operations.
Undo/rollback system for DSH: config & plugin snapshots, one-click undo/redo/restore, message-level revert, secret masking, safe mode, and offline WebUI/GUI/CLI rescue tools for when DSH won't start.