Skip to content
Back to skills

Lld Review

ASecurity

Review a low-level design (LLD): module and class structure, interfaces, data models, schemas, API contracts, sequence of calls and error handling, before or alongside implementation. Use when asked to review an implementation plan, interface design or data model.

  • 5 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 11, 2026
ai-agentsrustapi

Works with

  • api

Security analysis

A100/100

Scanned November 11, 2026

npx -y skills add thameema/memnos --skill lld-review --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Lld Review?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Lld Review
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/thameema-lld-review/badge)](https://www.skillsdirectory.com/skills/thameema-lld-review)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: lld-review
description: Review a low-level design (LLD): module and class structure, interfaces, data models, schemas, API contracts, sequence of calls and error handling, before or alongside implementation. Use when asked to review an implementation plan, interface design or data model.
---

# LLD review

Review the detailed design for correctness and fit. Read the existing code around the change first so you judge against the real conventions.

## Checks

1. **Interfaces.** Are signatures minimal, typed, and stable? Does each caller get what it needs without reaching into internals? Flag leaky abstractions, boolean-flag parameters that hide two behaviours, and public surface that nobody asked for.
2. **Data model.** Keys, nullability, constraints and indexes match the queries actually run. Check migrations are additive and reversible, and that old and new code can both run during a rolling deploy.
3. **State and concurrency.** Who owns each piece of state? Look for races, double-submit, non-idempotent retries, lock ordering, and transactions that span more than they should.
4. **Error handling.** Every external call has a timeout and a defined failure result. Errors carry enough context to debug and never swallow data loss. No catch-all that hides bugs.
5. **Validation at the boundary.** Untrusted input is validated once, where it enters; internal code can then trust it.
6. **Sequence.** Walk the main flow and each error flow call by call. Name the first step where an invariant could be violated.
7. **Reuse and consistency.** Does a helper, type or pattern for this already exist? Does naming and layering match the neighbours?
8. **Testability.** Can each rule be tested without a network or clock? List the 3–5 tests that would catch the most likely bugs in this design.

## Output

- **Verdict:** approve / approve with changes / rework.
- **Findings**, ranked, each with the file or section, what breaks and when, and the smallest fix.
- **Tests to add** (the 3–5 above).
- **What you did not check.**

Cite code, not taste. If something is merely different from how you would do it, leave it out.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.