More screenshots
Works with the AI tools you already use
Change Plan Reviewer
A rigorous reviewer that audits implementation plans for architectural risks, scope creep, and safety before you code.
Free
See it in action
You say
Review my plan to migrate users from the local auth table to Clerk. Objective: No downtime. Plan: 1. Install Clerk SDK. 2. Update AuthMiddleware. 3. Script to copy rows.
Your agent does
Verdict: REVISE BEFORE EXECUTION (Score: 65/100). Critical Risks:
- Missing dual-write phase for password sync during migration.
- No rollback plan for session invalidation.
- Sequential error: Script must run before Middleware update. Corrected plan included with specific file targets and test cases.
About this skill
The problem
Coding before planning leads to broken production data, circular dependencies, and massive PRs that are impossible to review. Developers often jump into implementation without considering backward compatibility, rollback procedures, or migration sequencing.
What it does
- Analyzes proposed file changes to identify missing dependencies or incorrect targets.
- Evaluates implementation sequences to prevent deployment-time failures in shared contracts or database migrations.
- Strips unrelated refactors and scope creep from the plan to keep PRs focused and safe.
- Defines mandatory test coverage and verification steps for each stage of the change.
- Provides a structured Ready/Revise/Do-not-execute verdict based on a 100-point quality rubric.
Why this beats prompting it yourself
Generic LLM prompts usually greenlight any plan that looks plausible. This skill applies a rigid risk catalog and execution-sequencing framework to catch subtle architectural violations, data integrity risks, and "point of no return" deployment flaws that simple prompts ignore.
Use cases
- Reviewing complex database schema migrations and data backfill strategies.
- Validating API contract changes between microservices to ensure backward compatibility.
- Refining high-risk infrastructure or authentication logic changes before touching a single file.
- Scoping feature implementations to ensure they don't leak logic across architectural boundaries.
Known limitations
This tool reviews plans but does not execute code, run tests, or modify files. It cannot verify findings perfectly if repository context (like full schema or architecture docs) is missing.
How to install
Drop the file into your AI Agent. Works with Claude, Cursor, ChatGPT, and 20+ more.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scanned
Verified clean today
- Free to download with an account