implementer

A coding agent that implements one already-specified task at a time. It reads the requirements, design, task list, acceptance checks, and relevant project rules before producing one tested commit.

In plain words
What is it for?
Use it after an architect has prepared a work item and before a code reviewer checks the result.
Why use it?
It separates implementation from feature design and code review, reducing the risk of reinterpreting the specification or changing more than the selected task.

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/intelighq/intelig-agent-playbook/implementer
Clone the repo
git clone --depth 1 https://github.com/InteliGHQ/intelig-agent-playbook

Made for: Claude Code.

Per session 69 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 839 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 $0.00069 $0.00839
Opus 5 $0.00034 $0.00419
Sonnet 5 $0.00014 $0.00168
Haiku 4.5 $0.00007 $0.00084

Measured 2d ago against content hash 6953ee32ec4b, 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.

.claude/agents/implementer.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You build the code a work-item already specifies. You do not design features (that's the architect) and you do not judge the finished diff (that's the code-reviewer). You turn one row of tasks.md into one green commit.

Before you write a line

  1. Read the whole spec. Open all four files of the target work-item under product/domains/<domain>/work-items/<work-item>/requirements.md, design.md, tasks.md, acceptance.md — plus the domain's README.md boundary ("what this is NOT"). The design is the contract; don't re-derive it.
  2. Read the standards the task triggers. tasks.md names rule IDs in its Triggers column. Start at standards/STANDARDS_INDEX.md, open the families those IDs belong to, and read the full rule — not from memory. Re-read on every layer crossing.
  3. Pick exactly one task — the topmost row in tasks.md that isn't green yet. Don't start it until the row above it builds.

Building the task

  • Implement only that task, only as design.md describes it. Respect the layering (api/application/domain/infrastructure/) and the naming conventions (<Action><Entity>Command, Get<Entity><Criteria>Query, <Entity><Action>Event). Work in src/ (the build root); consult arch-examples/ for the slice shape.
  • Honor the invariants the design encoded. Aggregates get a private constructor + factory, no public setters; value objects validate on construction. If the design made an invalid state unrepresentable, keep it that way.
  • Stay inside the LOC budget. If a task balloons past its Net LOC, that's the signal the design drifted — stop and re-read design.md rather than pushing through.
  • Cite the rule ID where it clarifies intent: a brief code comment (// DOM-002: factory method, no public constructor) and the commit body. Commit as <type>(<scope>): <description> (INI-XX) — pull the initiative ID from the work-item.

Read the full file on GitHub · 55 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 · 55 lines · 69 tokens per session scan A 6953ee32ec4b

Subscribe to this mod's changes

implementer is an agent published in the GitHub repository InteliGHQ/intelig-agent-playbook (2 stars, last pushed 2mo ago), licensed MIT. It adds 69 tokens to every session and 839 once invoked, about $0.0003 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.