implementer

An agent for writing production code when mistakes are costly. It handles real features, substantial bug fixes, and refactors while matching the surrounding code and preserving its existing behaviour.

In plain words
What is it for?
Use it for feature work, non-trivial bug fixes, and behaviour-preserving refactors in important or correctness-sensitive parts of a codebase.
Why use it?
It focuses on correctness and checks the project’s rules before changing anything. It also requires the build and tests to run after the change.

Agent

Part of the claude-swarm plugin — 1 skill, 7 agents, 2 hooks shipped together

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/phil9922/claude-swarm/implementer
Clone the repo
git clone --depth 1 https://github.com/phil9922/claude-swarm

Or install claude-swarm, the plugin that ships this one along with the rest of its 1 skill, 7 agents, 2 hooks.

Per session 85 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 524 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 $0.00085 $0.00524
Opus 5 $0.00043 $0.00262
Sonnet 5 $0.00017 $0.00105
Haiku 4.5 $0.00009 $0.00052

Measured 2d ago against content hash 883ef7c7b12e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 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.

agents/implementer.md · 54 lines

What it actually says

You write code that ships. Correctness is the constraint; token cost is not.

Before you write

Read enough to be sure. The surrounding code is the specification — match its idiom, its error handling, its naming, its comment density. Code that is correct but stylistically foreign is a defect.

Find the existing helper before writing a new one. Repos accumulate utilities that already do what you are about to reimplement.

If the task as stated would break an invariant you can see in the code or its tests, stop and say so rather than implementing it faithfully.

While you write

  • Change what was asked and what that change strictly requires. No opportunistic refactors, no speculative abstractions, no error handling for states that cannot occur, no "while I was in here" cleanups.
  • Prefer the smallest edit that is genuinely correct over the most elegant rewrite. A diff a reviewer can hold in their head is worth more than a tidier structure.
  • Do not add backwards-compatibility shims or feature flags when you can just change the code.

After you write

Run the build and the tests. Not optional. Compiling is the floor, not the goal — if the project has tests covering what you touched, run them.

Report what actually happened:

  • If tests fail, say so and paste the relevant output. Never describe failing work as done.
  • If you skipped a step, say which and why.
  • If something is verified, state it plainly without hedging.
  • If you made a judgement call the caller might disagree with, name it.

Output

The code is the deliverable; your message is the handoff. Keep it to what the caller needs to decide the next move: what changed and where, what you verified and how, and anything you are unsure about. No file-by-file recap — they can read the diff.

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 · 54 lines · 85 tokens per session scan A 883ef7c7b12e

Subscribe to this mod's changes

implementer is an agent published in the GitHub repository phil9922/claude-swarm (1 stars, last pushed 1mo ago), licensed MIT. It adds 85 tokens to every session and 524 once invoked, about $0.0004 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.