- Home
- Skills
- Code Quality & Review
- design review
Works with the AI Agents you already use
design review
Multi-lens expert audit that generates evidence-based security, architecture, and reliability scores for your codebase.
$39
design review
Example session with this skill installed
Run a design-review focusing on security and database lenses for the current repo. We've had some issues with slow queries and need to ensure our tenant isolation is airtight.
- Read your context and instructions
- Compiled the design review
I have initiated the design review using the
security and
database lenses. I've collected existing ADRs from docs/arch and identified 4 stateful components for CAP analysis. The reviewers are analyzing the schema and authorization matrices now.
Design Review Summary: Partial (2/6 Lenses)
| Lens | Score | Critical Findings | Top Risk |
|---|---|---|---|
| Database | 6.4/10 | 2 | Missing index on tenant_id in events table (F-DB-001) |
| Security | 4.8/10 | 3 | Broken Function Level Authorization in /api/admin (F-SEC-004) |
Overall Grade: 5.6/10
Executive Assessment
The system exhibits significant risks regarding data privacy. While the database schema is normalized, the lack of partitioned indexes for tenant_id leads to both performance degradation and potential cross-tenant data leakage during heavy query loads. The security lens identified that administrative endpoints rely on client-side checks rather than server-side role validation.
Critical Findings Matrix
F-DB-001 (High): Sequential scan on 40M row events table. Impact: Database CPU saturation during tenant exports.
F-SEC-004 (Critical): Direct Object Reference in UserManagement.ts. Impact: Any authenticated user can modify peer permissions by guessing a UUID.
F-SEC-007 (High): Secret leak in docker-compose.yaml. Impact: Production DB credentials stored in plaintext within the repository history.
Next steps
- Run
action-planto generate tickets for the critical security findings. - Review
docs/remediation/design-review-database.mdfor the full index optimization strategy. - Execute the
frontendlens to check if the admin UI exposes these vulnerable endpoints.
Connects securely to your tools. The creator never sees your data.
About this skill
Production-readiness is often a guessing game until the first major outage or security breach. code-review tools usually catch syntax errors or style violations but miss deep architectural flaws like race conditions, tenant isolation gaps, or CAP theorem misalignments. This skill provides an evidence-based design review through six expert lenses to identify structural risks before they become incidents.
What it does
- Multi-Lens Auditing runs specialized reviews for architecture, backend reliability, database integrity, frontend state, UX, and security.
- Severity-Ranked Reporting generates a unified remediation document with findings linked to specific files and line numbers.
- Quantified Scoring calculates a 0 to 10 grade for every system aspect using a strict rubric to measure production-readiness.
- Decision Tracking cross-references findings against existing Architecture Decision Records (ADRs) to ensure context-aware critiques.
- Security Coverage performs an audit including tenant isolation, authz matrices, and OWASP Top 10 coverage.
How it works
- Scope definition where you select specific lenses (e.g.,
security database) and provide context like known hotspots or incidents. - Context collection using
decisions.pyto aggregate ADRs and existing documentation so reviewers respect previous technical choices. - Parallel execution of up to three reviewer agents at a time, each producing a detailed lens report with evidence-backed findings.
- Final synthesis where
scorecard.pycomputes the arithmetic grades and a shell script merges reports into a singledesign-review.md.
Frameworks & tools
Uses Python 3 for scoring and decision collection scripts. Integrates with optional code-map and design-memory skills for enhanced repository traversal and cross-project precedent.
Why this beats prompting it yourself
Manual prompts often result in vague "best practice" advice that ignores your specific constraints. This skill enforces a strict quality gate, requiring concrete failure scenarios for critical findings and specific analysis for dual-writes, idempotency, and CAP analysis that generic prompts miss.
Use cases
- Pre-launch audit to find showstoppers in architecture or security before going live.
- Refactor planning to identify technical debt and coupling issues in a legacy monolith.
- Security hardening to map authorization flows and ensure OWASP Top 10 compliance.
- Reliability review to assess backend concurrency, idempotency, and data consistency.
Known limitations
Requires Python 3 for bundled scripts. Does not support automated remediation, focusing only on discovery and scoring. Frontend and UX lenses require visible client code or screenshots to be effective.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.
Requires Python 3. Does not perform auto-fixes. Effectiveness on UI/UX depends on available client code visibility. Scorecard script enforces strict grading caps that cannot be manually overridden.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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 12 days ago
- Software
Python