v1.0.0
First downloadable release of migration-guardian-agent.
Description
A read-only agent that reviews code and data migrations between two branches and raises evidence-backed alerts before merge, without touching the database or blocking the pipeline by default.
- 35 files
- Claude Code, Codex, ChatGPT, Cursor, GitLab CI, GitHub Actions
Read the start of every file
README.md
A read-only agent that reviews code and data migrations between two branches and raises evidence-backed alerts before merge, without touching the database or blocking the pipeline by default.
The rest of this file comes with the purchase.
agent-kits/INSTALL.md
Restart Claude Code: agents load at session start.
The rest of this file comes with the purchase.
agent-kits/_export/chatgpt/migration-guardian-agent/instructions.md
You are **migration-guardian-agent**. You protect code-and-data migrations before merge without changing the repository, database or merge request.
The rest of this file comes with the purchase.
agent-kits/_export/chatgpt/migration-guardian-agent/knowledge.md
Reviews code and data migrations between branches and raises evidence-backed alerts before merge—without touching the database or blocking the pipeline by default.
The rest of this file comes with the purchase.
agent-kits/_export/claude/agents/migration-guardian-agent.md
You are **migration-guardian-agent**. You protect code-and-data migrations before merge without changing the repository, database or merge request.
The rest of this file comes with the purchase.
agent-kits/_export/cursor/.cursor/commands/migration-guardian-agent.md
Reviews code and data migrations between two Git branches, reports destructive schema operations, applied-migration edits, unsafe backfills, blocking indexes and code/schema incompatibilities. Read-only and alert-only by default; GitLab and GitHub compatible.
The rest of this file comes with the purchase.
agent-kits/migration-guardian-agent/README.md
Reviews code and data migrations between branches and raises evidence-backed alerts before merge—without touching the database or blocking the pipeline by default.
The rest of this file comes with the purchase.
agent-kits/migration-guardian-agent/agent.md
You are **migration-guardian-agent**. You protect code-and-data migrations before merge without changing the repository, database or merge request.
The rest of this file comes with the purchase.
agent-kits/migration-guardian-agent/docs/customization.md
Create `.migration-guardian.json` at the client repository root from `config/config.example.json`.
The rest of this file comes with the purchase.
agent-kits/migration-guardian-agent/docs/integration.md
Copy the kit into the project or make it available in the runner, add `.migration-guardian.json`, then add this job:
The rest of this file comes with the purchase.
agent-kits/migration-guardian-agent/docs/usage.md
The scanner reads committed Git objects only. Fetch the base ref before running it. It does not inspect untracked files or connect to a database.
The rest of this file comes with the purchase.
agent-kits/migration-guardian-agent/templates/alert-summary.md
**Decision:** {{decision}} · **Highest severity:** {{highest_severity}}
The rest of this file comes with the purchase.
What is included
- Setup and usage guide
- 30 supporting files
- Licence, changelog and updates
Requirements
- Claude Code, or one of the bundled exports for Codex, ChatGPT or Cursor
- Python 3.9 or later — the scanner uses the standard library only
- Git, and a GitLab Merge Request or GitHub Actions job if you want it in the pipeline
Version history
Good questions, straight answers.
Does it connect to my database?+
No. It reads the migration files and the code of two Git refs; it never opens a database connection and needs no credential.
Will it block my merges?+
Not by default: it raises alerts and writes its report. A team can later opt into a failing threshold.
Which frameworks does it understand?+
Plain SQL files, and SQL statements inside migration files written in other languages such as PHP, Python or TypeScript; the paths are set per repository in its configuration.
Similar products
AI ToolkitTest Case Generator
An agent that turns requirements into positive, negative and edge test scenarios whose expected results come from the spec — and proves which rules are still untested.
View product$29.99
AI ToolkitSprint Planning Agent
An agent that proposes a sprint scope that fits the team's real capacity — in person-days, buffer kept, dependencies respected — and says exactly what was left out and why.
View product$29.99
AI ToolkitCI/CD Failure Analyzer
A read-only agent that finds why a pipeline failed — the first real error, its category, the probable cause, a fix and a ready-to-file bug ticket — with every claim tied to a log line or a commit.
View product$49.99
AI ToolkitRequirements Analysis Agent
An agent that makes a specification testable: every requirement atomic and traced, every vague word, contradiction and gap turned into a precise question for the business owner, before a line of code is written.
View product$29.99
AI ToolkitTechnical Documentation Generator
An agent that writes README, API and architecture docs that match the code as it is, and lists line by line the existing docs that lie: dead paths, removed scripts, unused variables. Drafts only, never overwrites.
View product$29.99
AI ToolkitTechnical Debt Analyzer
An agent that finds the debt that actually costs your team — files that change all the time, are big, untested and full of TODOs — ranks it with a visible formula and turns the top items into backlog entries.
View product$49.99
Cursor RulesProduction Telemetry and Live Debugging Rules
Four Cursor rules and a diagnostic playbook for answering a question about running production code without shipping a deploy to find out.
View product$29.99
AI ToolkitProject Health Monitor
An agent that gives an early, evidence-based health signal — schedule, delivery, quality, scope — rated GREEN, AMBER or RED against thresholds you set, with the trend since last time and the corrective actions.
View product$49.99