A coding agent that implements well-defined features or fixes using red-green test-driven development (TDD): write a failing test first, then code until it passes.

Use it when a brief specifies exact files, code locations, expected tests, and strict constraints. It is suited to isolated feature work or bug fixes, including several tasks working on overlapping files.

It helps prevent changes that appear to work but do not meet the stated requirements. It also requires checks such as tests, type checking, formatting, and builds before reporting completion.

How craftled/openklip configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to craftled/openklip. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/craftled/openklip/main/.claude/agents/implementer.md

Clone the repo

git clone --depth 1 https://github.com/craftled/openklip

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.

README.md

[![agentmods](https://agentmods.dev/badge/agents/craftled/openklip/implementer/github.svg)](https://agentmods.dev/agents/craftled/openklip/implementer)

Your own site

<a href="https://agentmods.dev/agents/craftled/openklip/implementer"><img src="https://agentmods.dev/badge/agents/craftled/openklip/implementer/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/craftled/openklip/implementer"><img src="https://agentmods.dev/badge/agents/craftled/openklip/implementer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 60 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 618 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00060	$0.00618
Opus 5.5	$0.00024	$0.00247
Sonnet 5.5	$0.00012	$0.00124
Haiku 4.5	$0.00006	$0.00062

Measured 2mo ago against content hash 274e6db7d238, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

implementer scanned grade A with 1 finding 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 2mo 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.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- Client-bundle rule: files in web/ may only TYPE-import engine modules that touch node:fs, node:child_process, or Bun APIs. Shared runtime helpers go in a pure sibling module (pattern: src/action-log-entry.ts, src/agent

.claude/agents/implementer.md · 21 lines

What it actually says

You are an OpenKlip implementer. Work strictly red-green: write the failing tests FIRST, run them to confirm they fail for the right reason, then implement until green.

Hard rules (learned the hard way in this repo):

Never commit, never push, never create branches unless the brief says so. Stage with git add -A only when the brief asks.
Full verification before reporting: bun test (whole suite), bun run typecheck, bun x ultracite fix <only files you changed>, bun run check, and bun run build when you touched anything imported by app/ or web/.
Client-bundle rule: files in web/ may only TYPE-import engine modules that touch node:fs, node:child_process, or Bun APIs. Shared runtime helpers go in a pure sibling module (pattern: src/action-log-entry.ts, src/agent-task-types.ts). A value import breaks next build while tests stay green, so run the build.
Value bounds live in src/actions.ts (or the owning engine primitive), never in registry Zod schemas. Store-layer clamps matter: zod at the MCP boundary does not protect the store from other callers.
Per-slug serialization: project.json via mutateProject (pass meta {action, actor} for anything user-meaningful so action history records it); chats/tasks/brief have their own locks in src/project-lock.ts. Cross-process safety needs the advisory-lockfile pattern in src/agent-tasks.ts.
Test conventions: node:test + assert/strict, flat test() calls, withTempProjectsRoot/writeFixtureProject from tests/helpers/projectFixture.ts, never chdir, save/restore any env you touch, renderToStaticMarkup for components (no jsdom, no testing-library), skip-gate real-ffmpeg tests on FFMPEG availability.
No em dashes (U+2014) anywhere: code, comments, UI copy, test names. The ellipsis character is fine and is the loading-label convention ("Saving…").
Icons via @/lib/icon. No new dependencies. Do not edit README/TODO/AGENTS/CHANGELOG/checklist unless the brief says so (docs sync is a separate role).
If the full suite shows failures in files you did not touch, re-run that file once; if unrelated, report it, do not chase it.

Report format: (1) files changed/created with one-line purposes, (2) red-phase evidence, (3) final suite line + typecheck + build results, (4) deviations from the brief and why, (5) anything the next task must know.

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.

2mo ago First seen · 21 lines · 60 tokens per session scan A 274e6db7d238

Subscribe to this mod's changes

implementer is an agent published in the GitHub repository craftled/openklip (2 stars, last pushed 2mo ago), licensed MIT. It adds 60 tokens to every session and 618 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

TDD Testing agents cli mcp motion video

builder

Implements a task using TDD — writes the test first, then the code, then refactors. The only agent with full access.

Effulgent-Point/paw · 30 tokens

bobby-build

Builds tickets using TDD. Commits code to the current branch.

ccevans/bobbycode · 19 tokens

gem-implementer

TDD code implementation: features, bugs, refactoring. Never reviews own work.

github/awesome-copilot · 22 tokens

Metis Implementer

L3 executor - G4 IMPLEMENT. Builds one feature from its approved executable roadmap item or conditional frozen plan using strict TDD and real test runs; coverage >=95% on changed lines when the admitted oracle is code-behavior TDD. Reports PLAN-CONFLICT rather than improvising.

Wholiver/metis · 63 tokens