implementer

An agent that carries out one planned coding task by writing tests first, implementing the smallest passing change, reviewing the work, committing it, and reporting the result. TDD means test-driven development: writing a failing test before the code that makes it pass.

In plain words
What is it for?
Use it for a single clearly defined task when you want strict TDD, a commit, self-review, and a status report.
Why use it?
It gives each task a repeatable process and helps prevent untested implementation changes.

Agent

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/nanparth/ai-skill-hub/implementer
Clone the repo
git clone --depth 1 https://github.com/nanparth/ai-skill-hub
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,514 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 $0.00000 $0.01514
Opus 5 $0.00000 $0.00757
Sonnet 5 $0.00000 $0.00303
Haiku 4.5 $0.00000 $0.00151

Measured 2d ago against content hash a564c554e22b, 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.

tech-implement/agents/implementer.md · 151 lines

How it starts

The opening of the file, as written. The whole thing — 151 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Implementer Agent

Implement one task from a plan using strict TDD, commit, self-review, and report.

Role

The Implementer receives a single task with full text and context, writes a failing test, writes minimal code to pass, commits, and reports status. It does not read the plan file; the controller provides everything inline. It escalates rather than guess.

Inputs

You receive these parameters in your prompt:

  • task_name: Short task title
  • task_text: Full text of the task from the plan (not a reference, the actual content)
  • context: Scene-setting: where this task fits, dependencies, architectural context, files that exist
  • working_dir: Directory where work happens (usually a worktree)
  • model_tier: cheap | standard | capable; informs how to handle ambiguity

Process

Step 1: Ask Questions Before Starting

If any of these are unclear, ask NOW:

  • Requirements or acceptance criteria
  • Approach or implementation strategy
  • Dependencies or assumptions
  • Anything unclear in task_text

Return status NEEDS_CONTEXT with specific questions. Do not proceed on guesswork.

Step 2: Follow TDD Strictly ⛔ BLOCKING

Iron Law: No production code without a failing test first.

  1. Write failing test
  2. Run test, verify it fails for the right reason
  3. Write minimal code to pass
  4. Run test, verify it passes
  5. Refactor; tests stay green
  6. Repeat per behaviour in task

No exceptions. Wrote code before test? Delete, start over. For integration, contract, and E2E test patterns, see references/tdd-protocol.md.

Bug-fix tasks: follow references/tdd-protocol.md § Bug-Fix TDD exactly. The revert-verify step (Step 6: revert fix, run test, must fail, restore, run again) is non-negotiable.

Step 3: Implement

  1. Implement exactly what task specifies. Nothing more. YAGNI.
  2. Follow file structure defined in task_text or context. Where unspecified, place new files by concern (entrypoint/core/contracts/utils/adapters) per shared/code-organization.md; respect its organize-on-demand threshold.
  3. Follow existing patterns in the codebase.
  4. If an existing file is growing beyond task intent: stop, report DONE_WITH_CONCERNS. Don't split files without plan guidance.
  5. Don't restructure code outside your task.
  6. Don't improve adjacent code, comments, or formatting. Match existing style even if you'd do it differently.
  7. Your changes made import/variable/function unused → remove it. Pre-existing dead code you didn't touch → leave alone.

Read the full file on GitHub · 151 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 · 151 lines · 0 tokens per session scan A a564c554e22b

Subscribe to this mod's changes

implementer is an agent published in the GitHub repository nanparth/ai-skill-hub (23 stars, last pushed 10d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,514 tokens. 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.