Provides a safe, documented workflow for executing DML (INSERT, UPDATE, DELETE) across many relational databases using the DbCli command-line tool. The skill includes concrete shell and programmatic examples, explicit backup-and-verify steps for UPDATE/DELETE, parameterized execution options, and recovery instructions. It returns structured JSON responses (e.g. {"AffectedRows": 5}) so agents can confirm results.
Use this skill when a user asks the agent to modify database data (insert rows, update fields, delete records) across supported RDBMS (SQLite, MySQL-family, PostgreSQL-family, SQL Server, Oracle, DB2, DaMeng, etc.). Always invoke the backup steps in the skill before running UPDATE or DELETE. Triggers: mentions of INSERT, UPDATE, DELETE, modify data, remove records, or execute DML.
Best for agents with shell and process execution capabilities (Copilot/Codex-style or Claude Code) and any agent that can set env vars and run CLI tools.
dbcli-exec is a documentation-only skill for executing DML (INSERT/UPDATE/DELETE) across 30+ databases via the DbCli CLI. It enforces a mandatory backup-before-destruction workflow with clear examples for SQLite, SQL Server, MySQL, PostgreSQL, Oracle, DB2, DaMeng, and GaussDB. No bundled scripts to test. Well-documented with Python and PowerShell wrappers, though programmatic examples use string interpolation that could enable SQL injection if table/column names aren't sanitized.
Solid safety-first documentation skill. The backup-before-destruction pattern is commendable. Main concern is the programmatic examples using string interpolation for SQL identifiers rather than parameterized approaches — though this is somewhat inherent to DDL-style operations. No scripts were present to test.