AgentSkillsCN

data-migration-verify

验证 PostgreSQL、MySQL 与 SQLite 数据库迁移的正确性。对比源与目标的表结构、表架构以及行数。适用于验证迁移过程或数据完整性时使用。

SKILL.md
--- frontmatter
name: data-migration-verify
description: Verify database migration correctness across PostgreSQL, MySQL, and SQLite. Compare source/target schemas, table structures, row counts. Use when validating migrations or data integrity.

Data Migration Verify

Verify database migration correctness. Supports PostgreSQL, MySQL, SQLite.

Inputs

InputTypeDefaultPurpose
db_typestringrequiredpostgres, mysql, or sqlite
source_dbstringrequiredSource connection string or path
target_dbstringrequiredTarget connection string or path
tablesstring""Comma-separated tables (all if empty)
migration_namestring""Migration being verified

Workflow

1. Load Persona

  • •persona_load("developer")

2. Check Known Issues

  • •check_known_issues("psql", "")
  • •check_known_issues("mysql", "")
  • •check_known_issues("sqlite", "")

3. PostgreSQL (if db_type == postgres)

  • •psql_schemas(connection="{{ source_db }}")
  • •psql_tables(connection="{{ source_db }}")
  • •psql_tables(connection="{{ target_db }}")
  • •psql_describe(connection="{{ source_db }}", table="{{ first_table }}") — if tables
  • •psql_size(connection="{{ source_db }}")
  • •psql_query(connection="{{ source_db }}", query="SELECT COUNT(*) FROM {{ first_table }}") — if tables
  • •psql_query(connection="{{ target_db }}", query="SELECT COUNT(*) FROM {{ first_table }}") — if tables

4. MySQL (if db_type == mysql)

  • •mysql_tables(connection="{{ source_db }}")
  • •mysql_tables(connection="{{ target_db }}")
  • •mysql_describe(connection="{{ source_db }}", table="{{ first_table }}") — if tables
  • •mysql_show_create_table(connection="{{ source_db }}", table="{{ first_table }}") — if tables
  • •mysql_query(connection="{{ source_db }}", query="SELECT COUNT(*) FROM {{ first_table }}") — if tables
  • •mysql_query(connection="{{ target_db }}", query="SELECT COUNT(*) FROM {{ first_table }}") — if tables

5. SQLite (if db_type == sqlite)

  • •sqlite_tables(database="{{ source_db }}")
  • •sqlite_tables(database="{{ target_db }}")
  • •sqlite_schema(database="{{ source_db }}")
  • •sqlite_describe(database="{{ source_db }}", table="{{ first_table }}") — if tables
  • •sqlite_query(database="{{ source_db }}", query="SELECT COUNT(*) FROM {{ first_table }}") — if tables
  • •sqlite_query(database="{{ target_db }}", query="SELECT COUNT(*) FROM {{ first_table }}") — if tables

6. Podman (optional)

  • •podman_exec(container="{{ source_db }}", command="echo 'database container accessible'") — if source is container

7. Failure Learning

  • •"connection refused" → learn_tool_fix("{{ db_type }}", "connection refused", "Database server not running", "Start DB or podman start <container>")
  • •"does not exist" → target DB not created

8. Memory

  • •memory_session_log("Data migration verify", "db_type={{ db_type }}, migration={{ migration_name }}")

Error Recovery

ErrorAction
"connection refused"Start database server, check connection params
"database does not exist"Create target database before migration

Output

Report: schemas, source/target tables, table descriptions, row counts, database size (Postgres), CREATE TABLE (MySQL), known issues. Compare source vs target for schema/row count mismatches.

Chains To

  • •workflow_health_check