unslop

Review changed code for AI-generated verbosity, unnecessary complexity, and defensive coding patterns — then apply concrete simplifications

Use this agent when reviewing code for AI-generated verbosity, unnecessary complexity, and defensive coding patterns. Examples: <example> Context: The user has just finished an AI code generation session and wants to clean up the output. user: "Clean up this AI-generated code, it's way too verbose" assistant: "I'll use the unslop-reviewer agent to evaluate the changed files against 14 simplicity directives — checking for unnecessary abstractions, defensive coding, over-split functions, and state bloat." <commentary> The user wants post-generation cleanup. The unslop-reviewer agent reads changed files, evaluates against directives, checks gotchas to prevent false positives, and outputs per-file findings with concrete fixes. </commentary> </example> <example> Context: The user notices over-engineered patterns in recently written code. user: "This function doesn't need to be this complicated, review the whole file" assistant: "I'll use the unslop-reviewer agent to analyze the file for over-engineering — unnecessary optional arguments, defensive null checks on non-nullable types, and functions split past readability." <commentary> The user spotted over-engineering. The agent systematically checks all 14 directives rather than just the one obvious issue, catching related problems the user may have missed. </commentary> </example> <example> Context: Pre-PR review to catch AI slop before it reaches code review. user: "Review my branch changes before I open a PR" assistant: "I'll use the unslop-reviewer agent to diff your branch against main and review every changed file for AI-generated verbosity, unnecessary state, and complexity that should be simplified." <commentary> Pre-PR quality gate. The agent diffs against the base branch, reviews only changed code, and produces a report suitable for guiding cleanup before the PR is opened. </commentary> </example>

This skill should be used when the user asks to "clean up", "remove slop", "simplify", "unslop", "review for quality", "reduce complexity", "tighten up", "cleanup before PR", "pre-PR review", "too verbose", "over-engineered", "remove unnecessary code", "AI generated mess", "simplify this", "reduce abstractions", or mentions "slop", "bloated code", "defensive coding", "unnecessary complexity", "too many functions", "over-abstracted". Also applicable when reviewing code after an AI generation session, preparing a branch for PR, or when code smells like it was written by a model that optimizes for looking thorough rather than being correct.