Database migration risk review
Skill alexzhu0/agent-ready-skills/skills/database-migration-risk-review
Ten practical AI-agent skills for turning messy work into clear context, evals, reviews, and launch-ready artifacts.
npx -y skills add alexzhu0/agent-ready-skills --skill database-migration-risk-reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when reviewing schema migrations, SQL changes, data backfills, index changes, or rollout plans for data loss risk, locking, rollback gaps, validation checks, and production safety.
SKILL.md
1.5 KB, as published. Nobody here has run it
Database Migration Risk Review
Purpose
Find practical risks in database migrations before they reach production.
Fit
- Use when schema, data, index, or backfill changes need production-safety review.
- Do not use when the task is general database modeling with no migration or rollout action.
Inputs
- Migration files, SQL snippets, schema diff, backfill plan, or rollout notes.
- Database type, table size, traffic patterns, and deployment process if available.
- Rollback expectations and validation commands.
Workflow
- Classify the migration type: schema, data, index, backfill, constraint, or cleanup.
- Identify locking, data loss, long-running query, and compatibility risks.
- Check forward and backward compatibility with application code.
- Recommend validation queries and rollout sequencing.
- State rollback limits honestly.
Output
Produce Markdown with:
- Migration Snapshot
- Risk Map
- Compatibility Review
- Validation Checks
- Rollout Plan
- Rollback Notes
- Open Questions
Validation
- Data loss risks are explicit.
- Locking and runtime impact are considered.
- Rollback is not promised when impossible.
- Validation checks are concrete.
- Unknown table size or traffic is flagged as a risk.