Code Review logo

Code Review

Community
dalemoncayo
code-review

Shared review rules for automated PR reviews and local manual reviews. Triggered by review commands.

Overview

Publisherdalemoncayo
Repositoryfullcase-web
Skill namecode-review
Stars
93
Forks
0
Bundled files
2
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by dalemoncayo on GitHub. Read the source before you install it.

Installation

Install the Code Review AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/dalemoncayo/fullcase-web/tree/main/.agents/skills/code-review
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/dalemoncayo/fullcase-web.git /tmp/fullcase-web
mkdir -p .claude/skills
cp -r /tmp/fullcase-web/.agents/skills/code-review .claude/skills/code-review
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Code Review in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Code Review on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Code Review is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Purpose

Review code changes in fullcase-web (Next.js 16 + TypeScript + Firebase client SDK) to detect bugs, design violations, spec violations, and security risks. Serves as the shared ruleset for both automated PR reviews (referenced by subagents) and local manual reviews (direct invocation).

Review Procedure

Step 1: Obtain the review target

Obtain the diff based on the provided arguments.

When a PR number is specified:

bash
gh pr diff <PR-number>

When a branch is specified:

bash
git diff <base-branch>...<target-branch>

If the base branch is omitted, default to main.

When nothing is specified (local review of own changes):

bash
git diff
git diff --cached
git status --short

If there are untracked new files, read the file contents directly for review.

For PR reviews (via the review-pr orchestrator): the diff is provided by the caller. Skip this step.

Step 2: Load project rules

Read AGENTS.md at the project root to understand project-specific rules. AGENTS.md is the canonical source of truth for the architecture, Firestore data model, folder responsibilities, and code-style rules.

Step 3: Execute review

Review the diff based on the "Review Criteria" below. If there are no findings, report "No issues found".

Step 4: Quality check

Before outputting findings, self-check against the following:

  • Does the finding match a pattern in references/bad-review.md?
  • Is it something ESLint / Prettier / tsc --noEmit would catch?
  • Is it a personal preference with no objective basis?

Exclude any findings that match.

Also reference .codex/prompts/codex-output-schema.json as a quality guide for finding structure (severity, priority, confidence, code location).

Step 5: Output

Output only findings that pass the quality check.


Review Criteria

Must detect

  1. Bugs / Logic errors — Unintended behavior, missed edge cases, unhandled null / undefined, missing await, stale closures, unsubscribed Firestore / Storage listeners
  2. Security — Hardcoded secrets, unsafe input, PII leaks, auth-guard bypasses, storage path abuse, open redirects, dangerouslySetInnerHTML without sanitization, firebase-admin on client
  3. Design violations — Firebase SDK imported directly into components, Server Actions / Route Handlers used for Firestore or Storage, types defined outside types/, manual edits to components/ui/, raw <img> instead of next/image, custom CSS in globals.css
  4. Firestore data-model violations — Dates stored as Date / ISO strings (must be serverTimestamp() / Timestamp), missing memberIds update on membership change, inconsistent collection paths vs. the canonical paths in AGENTS.md, non-atomic multi-doc writes where writeBatch / runTransaction is required
  5. Flow defects — (dashboard)/layout.tsx auth guard weakened, pendingInvite sessionStorage handling broken, invite token not validated before writing membership

Must NOT flag

  1. Issues detectable by linter / formatter / type-checker — Prettier, import sort, unused-imports, exhaustive-deps, any keyword alone, max-len
  2. Auto-generated / managed files — .next/, next-env.d.ts, components/ui/*, generated type declarations
  3. Naming preferences without basis — Do not flag if naming follows project conventions
  4. Code outside the diff — Unchanged code is out of scope
  5. Test coverage alone — Do not flag missing tests unless there is a concrete bug risk

Severity Definitions

LevelPriorityIconDefinitionExamples
CRITICALP0🔴Security vulnerability, data loss / leak risk, guaranteed production breakageSecret leak, auth guard bypass, dangerouslySetInnerHTML with raw user input, firebase-admin used on client
HIGHP1🔴Mandatory rule violation, definite bugFirebase SDK imported into a component, any used deliberately, date stored as ISO string, <img> instead of next/image, missing await on an async write
MEDIUMP2🟡Quality / performance improvementUnnecessary onSnapshot where getDocs suffices, missing loading/error in a hook, N+1 Firestore reads, inline object prop causing memoized-child re-renders
LOWP3🟢Minor improvement suggestionSmall refactor opportunity, non-critical naming polish, optional comment

Output Format

Each finding must be written in English.

🔴 [CRITICAL] `path/to/file.tsx:42`
Firebase SDK is imported directly inside a React component. Project rule: components consume hooks; hooks call services; only services touch the Firebase SDK.
Suggestion: Move the Firestore call into `services/<feature>-service.ts` and expose it via a hook under `hooks/`.

When there are no findings:

✅ No issues found

Project-Specific Checkpoints

Architecture & Data Access

  • Component → Hook → Service → Firestore SDK. Never skip a layer.
  • services/* contains pure async functions with zero React imports.
  • hooks/* always expose loading and error.
  • firebase-admin is forbidden. Server Actions / Route Handlers are forbidden for Firestore / Storage operations.
  • Types live in types/ and are imported via @/types.

Next.js 16 App Router

  • Route groups: (auth) public, (dashboard) protected, invite/[token] public.
  • (dashboard)/layout.tsx must be a client component and redirect via router.replace('/login') inside useEffect. Never use the server-only redirect() helper.
  • Use next/image everywhere; firebasestorage.googleapis.com and lh3.googleusercontent.com must be in next.config.ts remotePatterns.
  • Use next/font for fonts.

TypeScript

  • strict: true. Never any (explicit or implicit).
  • Use unknown + narrowing, Omit<T, 'id'|'createdAt'|'updatedAt'> for create payloads, Partial<Pick<T, ...>> for update payloads.

Firestore

  • Always serverTimestamp() or Timestamp for date fields. Never new Date() or ISO strings in stored values.
  • Use writeBatch for multi-document logical operations (e.g., invite accept).
  • Use setDoc(..., uid) — not addDoc — when a document ID must equal the user's UID.
  • Update projects/{id}.memberIds whenever a member is added/removed (security rules depend on it).

Forms & UI

  • Forms use react-hook-form + zod + shadcn Form primitives.
  • shadcn AlertDialog for all destructive confirmations.
  • Skeleton for loading states.
  • No CSS files, no CSS modules, no inline style props. No custom classes in globals.css.
  • components/ui/ is shadcn CLI output — never manually edit.

Error Handling

  • All async component actions wrapped in try/catch.
  • User-facing errors via toast.error(...); never raw Firebase error codes.
  • console.error only when process.env.NODE_ENV !== 'production'.
  • No console.log in committed code.

Security & PII

  • Firebase config only via process.env.NEXT_PUBLIC_FIREBASE_*.
  • No user emails / names / phone numbers in logs. Full UIDs too — prefer last-4-only (...ab12).
  • Storage paths must match the documented prefixes; enforce the documented MIME/size caps.
  • Validate invite tokens before writing members / memberIds.

Reference Examples

Refer to the following for review quality standards:

  • Good review examples: references/good-review.md
  • Anti-patterns to avoid: references/bad-review.md

Post-review feedback: If you discover missed issues or new pitfalls from this review, add them to the checklist via the update-review-checklist skill.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Code Review AI skill do?

Shared review rules for automated PR reviews and local manual reviews. Triggered by review commands.

Why use Code Review on TypingMind?

Because you install it once and use it with any model. Code Review is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Code Review in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/dalemoncayo/fullcase-web/tree/main/.agents/skills/code-review. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Code Review?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Code Review?

As many as you like. As long as a model supports skills, you can use Code Review with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Code Review AI skill free?

It is published on GitHub by dalemoncayo. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇