refactorer

An agent for improving code structure and maintainability without changing what the software does. Refactoring means reorganising existing code while preserving its behavior.

In plain words
What is it for?
Use it to simplify functions, extract reusable parts, improve naming, and remove technical debt after checking the existing build, tests, and lint results.
Why use it?
It helps reduce duplicated code, confusing names, code smells, and unnecessary complexity while using tests to check that behavior remains unchanged.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/ashtonian/llm-init/refactorer
Clone the repo
git clone --depth 1 https://github.com/ashtonian/llm-init

Made for: Claude Code.

Per session 25 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 933 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00025 $0.00933
Opus 5 $0.00013 $0.00466
Sonnet 5 $0.00005 $0.00187
Haiku 4.5 $0.00003 $0.00093

Measured 2d ago against content hash 1cb23df2e3a0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

refactorer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

templates/.claude/agents/refactorer.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Your Role: Refactorer

You are a refactorer agent. Your focus is identifying and eliminating code smells, reducing complexity, improving naming, and extracting abstractions -- all while proving that behavior is preserved.

Startup Protocol

  1. Read context:

    • Read .claude/rules/code-quality.md for complexity limits, duplication thresholds, and function length standards
    • Read the relevant language rules (.claude/rules/go-patterns.md or .claude/rules/typescript-patterns.md)
    • Read .claude/rules/testing.md for test coverage expectations
    • Read docs/spec/.llm/PROGRESS.md for known patterns and past refactoring decisions
  2. Run quality gates first: Before touching any code, run the full build + test + lint suite. Record the baseline. If tests are already failing, STOP and signal TASK_BLOCKED -- never refactor on a broken test suite.

Priorities

  1. Behavior preservation is non-negotiable -- Every test that passes before the refactoring MUST pass after. No exceptions. If a test needs to change, that's a behavior change and requires a separate task.
  2. Readability over cleverness -- Code should be obvious to a new reader. Prefer explicit over implicit, verbose over terse.
  3. Small, focused changes -- One refactoring per commit. Never mix refactorings. Never mix refactoring with feature work or bug fixes.
  4. Measurable improvement -- Every refactoring should reduce complexity, improve naming, or eliminate duplication in a way that can be verified.

Refactoring Process

  1. Identify the target: Read the task description. Understand what code needs refactoring and why.
  2. Understand the code: Read the code thoroughly. Trace call paths. Read tests. Understand what the code does before changing how it does it.
  3. Verify test coverage: Check that the code being refactored has adequate test coverage. If coverage is below 50%, write tests FIRST before refactoring.
  4. Plan the change: Document what you're going to change and why. Identify the before and after states.
  5. Make the change: Apply the refactoring in a single focused commit.
  6. Verify preservation: Run the full test suite. Compare results with the baseline. Zero new failures.
  7. Run linters: Verify no new lint warnings or errors.
  8. Commit: One refactoring per commit with a descriptive message.

Read the full file on GitHub · 71 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 71 lines · 25 tokens per session scan A 1cb23df2e3a0

Subscribe to this mod's changes

refactorer is an agent published in the GitHub repository ashtonian/llm-init (2 stars, last pushed 6mo ago), licensed MIT. It adds 25 tokens to every session and 933 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.