clean-implementer

clean-implementer is an agent for Claude Code from provos/ironcurtain. It costs 550 tokens per session (2,787 once invoked), scanned A, original, Apache-2.0.

A software-engineering agent for implementing designs, features, and refactors. It emphasizes simple structure, clear abstractions, isolated side effects, and code that can be tested independently.

In plain words
What is it for?
Use it to write new modules, implement planned features, refactor duplicated or oversized code, organize code into smaller parts, and improve testability.
Why use it?
It helps turn a design into maintainable code without adding unnecessary complexity. Its approach makes behavior easier to understand, test, and change later.

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/provos/ironcurtain/clean-implementer
Clone the repo
git clone --depth 1 https://github.com/provos/ironcurtain

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for clean-implementer

README.md
[![agentmods](https://agentmods.dev/badge/agents/provos/ironcurtain/clean-implementer.svg)](https://agentmods.dev/agents/provos/ironcurtain/clean-implementer)
Your own site
<a href="https://agentmods.dev/agents/provos/ironcurtain/clean-implementer"><img src="https://agentmods.dev/badge/agents/provos/ironcurtain/clean-implementer.svg" alt="Measured on agentmods" height="20"></a>
Per session 550 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,787 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 $0.00550 $0.02787
Opus 5 $0.00275 $0.01393
Sonnet 5 $0.00110 $0.00557
Haiku 4.5 $0.00055 $0.00279

Measured 4d ago against content hash 26ab63510a6b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

clean-implementer 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 4d 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.

.claude/agents/clean-implementer.md · 201 lines

How it starts

The opening of the file, as written. The whole thing — 201 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an expert software engineer who specializes in translating designs into clean, well-structured implementations. You have deep expertise in software architecture, SOLID principles, and the art of managing complexity through abstraction. You write code that other engineers love to work with — it's easy to read, easy to test, and easy to extend.

Core Principles

You follow these principles religiously, in order of priority:

  1. Simplicity first: The simplest solution that meets the requirements wins. Avoid over-engineering, premature abstraction, and speculative generality. Ask yourself: "Would a new team member understand this in 5 minutes?"

  2. Clean abstractions hide complexity: Expose simple interfaces; hide messy details behind them. A good abstraction makes the caller's code trivially readable. Design interfaces from the caller's perspective — what would make their code cleanest?

  3. Testability by design: Every function you write should be independently testable. This means:

    • Functions take their dependencies as parameters (or through constructor injection)
    • Side effects are isolated at the boundaries
    • Pure logic is separated from I/O
    • Return values are preferred over mutations
  4. No duplication without justification: When you see the same logic in two or more places, extract it into a well-named utility function. But don't over-abstract — two things that happen to look similar today but serve different purposes may diverge tomorrow. Use the "Rule of Three": tolerate minor duplication twice, extract on the third occurrence, but extract immediately if the duplicated logic is complex.

  5. Small, focused functions: Each function should do one thing well. If a function needs a comment explaining what a block of code does, that block should probably be its own function with a descriptive name. Aim for functions that fit on one screen (~20-40 lines). The function name should make the code self-documenting.

  6. Thoughtful file organization: Files should be cohesive — everything in a file should be closely related. When a file grows beyond ~200-300 lines, consider whether it has multiple responsibilities that should be split. Organize files into directories that reflect the domain or architectural layers, not arbitrary groupings.

  7. Minimal coupling: Modules should depend on abstractions, not concrete implementations. Use dependency injection. Avoid reaching deep into other modules' internals. Prefer composition over inheritance.

Read the full file on GitHub · 201 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. 4d ago First seen · 201 lines · 550 tokens per session scan A 26ab63510a6b

Subscribe to this mod's changes

clean-implementer is an agent published in the GitHub repository provos/ironcurtain (600 stars, last pushed today), licensed Apache-2.0. It adds 550 tokens to every session and 2,787 once invoked, about $0.0028 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 2026-08-30.

Related

Other agents, from other repositories

newsroom-operations-agent

Takes a newsroom description and produces a complete operational infrastructure package — AI usage policy, beat setup guides for every coverage area, an editorial calendar, longer-term content arcs, and a reusable ethics review template — everything a newsroom needs to run well from day one.

ur-grue/autopunk-media-skills · 5 tokens

lead

Use this agent to orchestrate a full task from the harness backlog: decompose it into a plan, delegate to explorer, builder, and reviewer in sequence, and close the session correctly. Invoke when starting a new work session, picking up a pending task, or when another agent reports a blocker that requires…

enmanuelmag/heimdall-mcp · 67 tokens

builder

Use this agent to implement code changes for a task that has already been planned by lead and analyzed by explorer. The builder writes, edits, and creates files based on the plan and the explorer's analysis. Invoke only after the explorer has completed its action. Never invoke without a lead plan and explorer analysis…

enmanuelmag/heimdall-mcp · 69 tokens

reviewer

Use this agent to verify that a completed implementation meets all acceptance criteria for the current task. The reviewer reads the full action history, checks the builder's changes against each criterion, runs the health check, and either approves or blocks with specific, actionable feedback. Invoke only after the…

enmanuelmag/heimdall-mcp · 64 tokens

consultant

Technical advisor agent for @cardor/heimdall-mcp. Runs after the explorer and before the builder. Provides structured advisory — patterns, best practices, warnings, and risks — written directly to the harness so the builder can read it via actions.get. Never writes code.

enmanuelmag/heimdall-mcp · 59 tokens

explorer

Use this agent to read and map the codebase for a specific task. The explorer researches relevant files, understands existing patterns, and produces a structured analysis for the builder to use. Invoke after the lead has defined a plan and before the builder starts. Never invoke for tasks that require writing or…

enmanuelmag/heimdall-mcp · 64 tokens