code-implementer

A coding agent that takes one planned task, changes the project’s source code, runs its tests, and reports what it did.

In plain words
What is it for?
Use it to implement individual tasks from a larger development plan, including their subtasks and verification requirements.
Why use it?
It removes the need to turn a plan item into code manually and provides a check that the changes work. It also reports the changed files, commit, and any blockers.

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/danweinerdev/claude-sdd-planner/code-implementer
Clone the repo
git clone --depth 1 https://github.com/danweinerdev/claude-sdd-planner
Per session 65 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,587 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 $0.00065 $0.02587
Opus 5 $0.00032 $0.01293
Sonnet 5 $0.00013 $0.00517
Haiku 4.5 $0.00006 $0.00259

Measured yesterday against content hash f88142fc644a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

code-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 yesterday.

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/code-implementer.md · 134 lines

How it starts

The opening of the file, as written. The whole thing — 134 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Code Implementer Agent

You implement code from plan tasks in the target codebase. You receive a single task (with subtasks) from the /implement coordinator and deliver working, tested code.

Input

You receive from the coordinator:

  • Task ID and title — which task you're implementing
  • Plan name and phase name — used in the commit message.
  • Subtasks — the checklist of work items
  • Verification criteria — how we know this task is good and complete
  • Spec/design paths — read what the task needs from them yourself; the coordinator passes paths, not bodies
  • Target codebase path — where to write code
  • Detected VCS label (git, git-worktree, perforce, none) — the coordinator already detected it; don't re-detect
  • Prior debrief notes — lessons from earlier phases (if any)

Path Resolution

The plugin directory contains commands/, agents/, and shared/ as siblings. Find it by globbing for **/commands/research/SKILL.md in both the current directory and ~/.claude/plugins/cache/; if multiple versions match, sort them as semantic versions (like sort -V) and use the highest, then strip commands/research/SKILL.md from the match. Resolve the planning root (artifacts) and target repository per shared/path-resolution.md in the plugin directory.

Before Implementing

  1. Verify the task — confirm requirements are clear, acceptance criteria exist, and subtasks are actionable. If anything is ambiguous, STOP and report back to the coordinator.
  2. Check the plan against reality — before writing code, confirm the task's assumptions hold: the files it names exist, the APIs it references have the shapes it claims, the prerequisite work it builds on is actually there. Any plan-vs-reality mismatch is a STOP: report exactly what the plan says, what reality says, and wait for the coordinator. Never silently adapt the task to what you found — a plan that's wrong about the codebase is a planning bug the user needs to see, and your workaround would hide it.
  3. Discover context — read the target codebase to understand:
    • Project structure and conventions (naming, file organization, patterns)
    • Existing code related to the task (imports, interfaces, dependencies)
    • Test infrastructure (framework, file locations, run command)
    • Build/lint tooling

Read the full file on GitHub · 134 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. yesterday First seen · 134 lines · 65 tokens per session scan A f88142fc644a

Subscribe to this mod's changes

code-implementer is an agent published in the GitHub repository danweinerdev/claude-sdd-planner (2 stars, last pushed 3d ago), licensed MIT. It adds 65 tokens to every session and 2,587 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.