Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Code Reviewer

ASecurity

Use this skill when reviewing any code — general PRs or AI/LLM-specific pipelines. Triggers include: "review this code", "check my PR", "find issues", "is this safe", or when shown code that interacts with LLM APIs, prompt construction, vector stores, agents, or any backend/frontend logic.

3 stars
0 votes
0 copies
0 views
Added 9/19/2026
ai-agentsgosqltestingcode-reviewapidatabasefrontendbackendsecurityperformance

Works with

api

Security Analysis

A100/100

Scanned 9/19/2026

Install to Claude Code

$npx -y skills add satishkc7/claude-config --skill code-reviewer --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Code Reviewer?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Code Reviewer
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/satishkc7-code-reviewer/badge)](https://www.skillsdirectory.com/skills/satishkc7-code-reviewer)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: code-reviewer
description: Use this skill when reviewing any code — general PRs or AI/LLM-specific pipelines. Triggers include: "review this code", "check my PR", "find issues", "is this safe", or when shown code that interacts with LLM APIs, prompt construction, vector stores, agents, or any backend/frontend logic.
---

# Code Reviewer Skill

Review code like a mentor, not a gatekeeper. Every comment should teach something. Be specific, explain why, prioritize ruthlessly.

## Severity System

- **BLOCKER** — must fix before merge (data loss, security vulnerability, broken functionality)
- **SUGGESTION** — should fix soon (missing validation, performance issue, unclear logic)
- **NIT** — nice to have (style, minor naming, docs)

---

## Pass 1: AI/LLM-Specific Issues (run if code touches LLM APIs, prompts, agents, or vector stores)

### BLOCKER
- Prompt injection: user input injected directly into prompt string
  - BAD:  `prompt = f"Summarize: {user_input}"`
  - GOOD: `prompt = f"Summarize:\n<document>{sanitize(user_input)}</document>"`
- Hardcoded vague model version (use full pinned version string)
- Missing `max_tokens` on any API call
- No retry/backoff logic on LLM calls
- PII or secrets logged in prompt debug output
- API keys in prompt text or hardcoded in source

### SUGGESTION
- Concatenated prompt strings instead of template functions
- No output validation or JSON parse error handling
- Unbounded conversation history (no token budget management)
- LLM calls not abstracted behind a single wrapper
- No token usage or cost tracking per call

### BEST PRACTICE
- Prompts versioned separately from business logic
- Agent tool calls have input validation and output schemas
- Streaming errors handled (not just final response errors)
- Model fallback defined if primary model unavailable

---

## Pass 2: Security (all code)

### BLOCKER
- SQL/NoSQL injection (user input in raw queries)
- XSS (unsanitized user input rendered as HTML)
- Auth bypass (missing auth checks on protected routes)
- Insecure direct object reference (no ownership check on resource access)
- Secrets hardcoded in source (API keys, passwords, tokens)
- Missing input validation on any public endpoint

### SUGGESTION
- No rate limiting on public-facing endpoints
- Error messages exposing stack traces or internal paths to users
- Missing CSRF protection on state-changing requests
- Passwords compared without constant-time function

---

## Pass 3: Correctness & Logic

### BLOCKER
- Race conditions or deadlocks
- Off-by-one errors in critical loops
- Null/undefined dereference without guard
- Breaking change to a public API contract

### SUGGESTION
- Missing error handling on async operations
- Silent catch blocks (`catch (e) {}`)
- Edge cases not handled (empty array, zero, negative numbers)
- Business logic doesn't match requirements

---

## Pass 4: Performance

### SUGGESTION
- N+1 queries (database call inside a loop)
- Synchronous blocking call in async context
- Unbounded list fetched when paginated would suffice
- Expensive computation run on every render/request instead of cached

---

## Pass 5: Maintainability

### SUGGESTION
- Functions > 40 lines doing more than one thing
- Magic numbers without named constants
- Variable names that don't describe intent (`data`, `temp`, `obj`)
- Deeply nested conditionals (> 3 levels) — use early return

### NIT
- Dead code or commented-out blocks
- Unused imports or dependencies
- Inconsistent naming conventions
- Missing docstring on public functions with non-obvious behavior

---

## Pass 6: Testing

### SUGGESTION
- No tests for core business logic
- Happy path only — no edge case or error path coverage
- Tests that only test implementation (not behavior)
- Flaky tests (time-dependent, order-dependent)

---

## Standard Engineering Checklist

- [ ] Error handling on all exceptions
- [ ] Type hints/annotations present
- [ ] Single-responsibility functions
- [ ] Named constants (no magic numbers)
- [ ] Tests for core logic
- [ ] Dependencies pinned
- [ ] No dead code or debug artifacts
- [ ] No secrets in source

---

## Output Format

```
## Summary
[2-3 sentence overview: what the code does, overall quality, key concerns]

## BLOCKERS
[List each — location, what it is, why it matters, how to fix]

## SUGGESTIONS
[List each — location, what it is, reasoning, suggested fix]

## NITS
[Brief list — location, what to tweak]

## What's Working Well
[Specific callouts of good patterns worth replicating]

## Decision
APPROVED | APPROVED WITH SUGGESTIONS | NEEDS CHANGES
```

**Rules:**
- Be specific ("line 42 SQL injection via f-string" not "security issue")
- Always explain the why, not just the what
- Praise good code — call out clever solutions and clean patterns
- One complete review — don't drip-feed feedback across rounds
- Never block on style if a linter handles it

Attribution

satishkc7satishkc7
View sourceMore from satishkc7 →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

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

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.

1023331 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

686011 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3331 votes

catchup

Recovers prior coding-agent session context by running `catchup <agent> --since-compact`, which extracts a clean summary of a previous Codex, Claude Code, Antigravity, OpenCode, or Pi Agent session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", or asks to recover/summarize a previous session before continuing. Do NOT use for the current conversation, git history, or any non-agent log.

611 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →