dsh-local-dba
Run the following command in DeepSeek Harness:
dsh plugin install kichare/dsh-local-dba
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install kichare/dsh-local-dba in DeepSeek Harness to install this plugin (source: https://github.com/kichare/dsh-local-dba). Restart the profile after installation to load the bundle.
About this plugin
Working with a database inside a DeepSeek Harness workflow used to mean leaving the chat, opening a terminal, and typing mysql or psql by hand. dsh-local-dba folds that ritual into the agent itself: once the plugin bundle is registered, a full set of db_* tools lets the agent run queries, inspect schemas, execute DDL/DML, perform logical backups and restores, and pull slow queries or process lists—all against MySQL / MariaDB and PostgreSQL.
Safety is baked in by design. db_query is strictly read-only and enforced at the SQL level; the write-permission and backup/restore switches are both off by default, so the plugin is safe the moment it loads. Passwords can come from an environment variable, and credential values are passed to mysqldump / pg_dump child processes via MYSQL_PWD and PGPASSWORD so they never appear in the process argument list. Managing multiple connections is a UI exercise rather than a file-editing chore: add, remove, or set the default connection in the plugin settings card, and changes hot-reload without a restart. The whole bundle is plain ESM JavaScript with no build step.
It is aimed at solo developers juggling several local databases, full-stack engineers who want to weave SQL steps into an agent pipeline, and DBAs who would rather delegate repetitive tasks like schema inspection, slow-log triage, and quick backups. When you do flip on write access, the agent can issue CREATE / ALTER / DROP / INSERT / UPDATE / DELETE statements—changes are irreversible, so keep those switches off for production or critical instances and enable them only in isolated local or test profiles.
Use Cases
- Query a local database and inspect tables, indexes, and foreign keys directly from an agent chat session
- Have the agent pull slow queries and process lists to help isolate performance bottlenecks
- Run logical backup and restore steps inside an agent pipeline instead of switching to a terminal
Best For
- Solo developers juggling several local databases
- Full-stack engineers embedding SQL steps into an agent pipeline
- Ops engineers or DBAs needing quick schema checks and slow-log triage
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.