Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    DB Migration Safety Reviewer

    1

    Review schema migrations for lock risk, data loss, backfill cost, and rolling-deploy breakage — get a verdict, safe rewrite, and deploy sequence.

    $15

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    DB Migration Safety Reviewer

    DB Migration Safety Reviewer

    Example session with this skill installed

    Database: PostgreSQL 15
    Table size: not provided (production; estimate unknown)
    Deployment: rolling deploy (no maintenance window)
    Context: billing service; subscriptions table is hot path

    Migration SQL

    -- migration_2026_03_12_subscriptions.sql

    ALTER TABLE subscriptions
    ADD COLUMN plan_tier TEXT NOT NULL;

    CREATE INDEX idx_subscriptions_account_id

      ON subscriptions(account_id);
    
    UPDATE subscriptions
    SET trial_ends_at = created_at + INTERVAL '14 days'
    WHERE trial_ends_at IS NULL;
    

    ALTER TABLE subscriptions
    DROP COLUMN old_price_cents;

    • Read your context and instructions
    • Compiled the db migration safety

    DB Migration Safety Review

    1. Risk Summary

    | Database | PostgreSQL 15 |
    | Statements reviewed | 4 |
    | Migration Risk Verdict | BLOCK |
    | Risk Score | 0 / 100 |
    | Critical / High / Medium / Low | 1 / 3 / 0 / 0 |
    | Table size assumption | large (not provided) |

    3. Risk Findings

    MIG-01 CRITICAL lock — ADD plan_tier NOT NULL without DEFAULT
    → Add nullable/DEFAULT, backfill, then SET NOT NULL
    MIG-02 HIGH lock — CREATE INDEX blocks writes
    → CREATE INDEX CONCURRENTLY + verify valid
    MIG-03 HIGH backfill — Unbatched UPDATE on hot table
    → Batch 1k–10k with statement_timeout
    MIG-04 HIGH data-loss — DROP old_price_cents irreversible
    → Stop reading in code first; drop later

    6. Deploy Sequence

    Expand → Migrate → Contract (destructive last)

    JSON: verdict "BLOCK", risk_score 0
    Disclaimer: Review aid; not a substitute for staging tests, backups, or DBA review.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    The problem

    Schema migrations are the quietest way to cause an outage: table locks, failed NOT NULL adds, irreversible drops, and rolling-deploy breaks when old pods still read a column you just removed. Generic “review this SQL” prompts miss database-specific lock behavior and rarely hand you a safe rewrite.

    What it does

    DB Migration Safety Reviewer is a production-safety review for DDL/DML migrations. It returns:

    • A Migration Risk Verdict: SAFE-TO-SHIP / SHIP-WITH-CAUTION / BLOCK
    • A deterministic Risk Score (0–100)
    • Severity-ranked findings across lock, data-loss, backfill, compatibility, and rollback
    • A safe-execution rewrite for every risky statement (e.g. CONCURRENTLY, batched backfill, expand-migrate-contract)
    • Rollback plan + rolling-deploy sequence
    • Pre-flight checklist + Migration Review JSON for CI/docs

    Frameworks & tools

    • PostgreSQL, MySQL, and SQLite behavior differences
    • Five-dimension risk model + hard verdict gates
    • Common patterns: indexes, type changes, NOT NULL, FK validation, drops, large UPDATEs
    • Works from pasted SQL or migration files in your agent

    Why this beats prompting it yourself

    • Database-aware rules (e.g. Postgres CONCURRENTLY vs MySQL online DDL tools)
    • Hard gates so CRITICAL issues force BLOCK — not a soft “maybe fine”
    • Safe rewrites, not only warnings
    • Rolling-deploy compatibility built into the plan
    • Assumes large tables when size is unknown (worst case) and labels assumptions
    • Read-only: never executes against your database

    Use cases

    • Backend engineers shipping migrations under rolling deploys
    • DBAs reviewing lock and data-loss risk before a release
    • Teams adding indexes without a maintenance window
    • Destructive or backfill-heavy changes that need a staged plan

    Known limitations

    • Cannot measure real lock duration or row counts without that input
    • Version-sensitive advice is sharper when you state the DB version
    • Not a substitute for staging tests, backups, or DBA review on high-risk production changes
    • Does not connect to or execute against any database

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Data & Databases.

    • Cannot measure real lock duration or table size without input (assumes large when unknown).
    • Advice is version-sensitive — provide PostgreSQL/MySQL/SQLite version when possible.
    • Review aid only — not a substitute for staging tests, backups, or DBA review.
    • Never connects to or executes against a database.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 20 days ago

    Permissions

    Read Files

    Read-only review. Does not execute DDL/DML, open DB connections, or apply changes. If auto-detect grants Shell/Write/Network, turn them off.

    Listed3 months ago

    What's inside

    Frequently Asked Questions

    The skill identifies high-risk operations like table locks, irreversible data loss, and breaking changes for rolling deployments. It provides a risk verdict (0-100), a severity-ranked list of findings, and a 'safe-execution' rewrite for every risky statement.

    It supports PostgreSQL, MySQL, and SQLite. For the most accurate advice, it is recommended to specify your database version so the agent can account for specific engine behaviors like Postgres CONCURRENTLY or MySQL online DDL tools.

    No, the skill is read-only and never connects to or executes code against your database. It analyzes SQL scripts or migration files you provide and returns a safe execution plan for you to apply manually or through your CI/CD pipeline.

    Absolutely. The reviewer checks if a migration will break active sessions (such as removing a column still being read by old pods) and provides an 'expand-migrate-contract' sequence to ensure zero-downtime deployments.

    If you don't provide table sizes, the skill defaults to a 'worst-case' assumption, treating tables as large and mission-critical. It will explicitly label these assumptions in the report and suggest strategies like batched backfills and index creation with low lock priority.

    You receive the full skill logic compatible with your AI agent, a pre-flight checklist for deployments, and a Migration Review JSON template that can be integrated into your team's documentation or CI logs.

    New in Data & Databases