無料GitHub で公開

review-ai-slop

Review code for AI slop patterns. TRIGGER when user says "review for AI slop", "check for AI patterns", "clean up AI code", "audit AI-generated code", or "review AI slop".

archibate/dotfiles-opencode

インストール方法を見る

含まれるファイル（1）

SKILL.md3.8 KB

SKILL.md（原文）

インストールする前に、エージェントに与えられる指示の中身を確認できます。

AI Slop Review Checklist

Review the specified code for AI-generated slop patterns.

Target: $ARGUMENTS (or current git diff if no arguments provided)

Output format: Group findings by category. For each: file path, line number, pattern name, brief description, suggested fix. Do not make inline edits — report findings and let user decide what to fix.

A. Defensive Programming

Load the anti-defensive skill and apply all 10 patterns:

Swallowing exceptions
Dictionary defaults on required fields
Null coalescing to fabricate data
Type coercion instead of validation
Compatibility shims
Unnecessary null checks
Catch-all exception handlers
Over-validation at internal boundaries
Fabricated default values
Logging warnings instead of raising

B. Over-Engineering

Pattern	Flag When
Premature abstraction	Factory/strategy/abstract base class for a single implementation. "In case we need it later."
Unnecessary wrappers	`get_config()` that just calls `os.environ.get()`. Utility functions invoked only once.
Over-modularization	Simple logic split across many files when one suffices. Deep directory nesting for small codebase.

C. Unnecessary Infrastructure

Pattern	Flag When
Retry logic everywhere	Exponential backoff on local function calls or non-transient operations.
Config/env var overuse	Every constant configurable via env var, even ones that should never change (math constants, format strings).
Unnecessary async	asyncio/threading on CPU-bound or already-fast operations (<10ms).
Feature flags for no reason	Boolean toggles controlling behavior that has no reason to vary.

D. Code Bloat

Pattern	Flag When
Excessive logging	`logger.info` on every function entry/exit. Logging parameter values and return values of internal functions.
Magic number extraction	`MAX_SIZE = 100` when `100` is obvious from context. Constants that add no clarity.
Trivial docstrings	`def add(a, b): """Add a and b."""` — docstring restates function name.
Unnecessary data classes	dataclass/TypedDict created for one-off use instead of just using a dict or tuple.

E. Testing Slop

Pattern	Flag When
Tests that test nothing	Asserting a function exists or returns the right type, not that it returns correct values. Tests with no assertions.
Over-mocking	Mocking so heavily the test only validates the mock, not the real logic. Mocking what you're trying to test.

F. Other Patterns

Pattern	Flag When
Callback/event overuse	Event emitters, hook systems, or pub/sub for simple linear flows.
Runtime type checking	`isinstance` checks inside functions that duplicate what type hints + static analysis already handle.
Hallucinated imports	Using APIs from wrong library version. Importing packages that don't exist.
Boilerplate generators	Code that exists to satisfy a template rather than solve a problem. Empty `__init__.py` with docstrings.

Summary Table

Category	Count	Source
A. Defensive Programming	10	`anti-defensive` skill
B. Over-Engineering	3	This skill
C. Unnecessary Infrastructure	4	This skill
D. Code Bloat	4	This skill
E. Testing Slop	2	This skill
F. Other	4	This skill
Total	27

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

ログインしてレビューを書く

同じリポジトリのスキル

概要と使いどころ

agent-browser

無料

Browser automation CLI for AI agents. Use when needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. TRIGGER when user requests to "open a website", "fill out a form", "click a button", "take a screenshot", "debug this in browser", "scrape data from a page", "test this web app", "login to a site", "frontend UI/UX aesthetics", "automate browser actions", or any task requiring programmatic web interaction.

日本語の概要は準備中です。原文の説明を表示しています。