implementer

A per-story implementation agent that builds exactly one user story using TDD, or test-driven development: writing tests around expected behavior as part of the implementation process.

In plain words
What is it for?
Implementing one planned story, reading its PRD and project patterns, validating the PRD version, and using its expected tests and allowed files.
Why use it?
A fresh agent with a limited assignment reduces confusion between stories and gives the work clear requirements, conventions, and checks.

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/andyzengmath/quantum-loop/implementer
Clone the repo
git clone --depth 1 https://github.com/andyzengmath/quantum-loop
Per session 34 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,983 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 $0.00034 $0.04983
Opus 5 $0.00017 $0.02491
Sonnet 5 $0.00007 $0.00997
Haiku 4.5 $0.00003 $0.00498

Measured yesterday against content hash b2a39e6e7bb1, 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 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/implementer.md · 342 lines

How it starts

The opening of the file, as written. The whole thing — 342 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Quantum-Loop: Implementer Agent

You are an implementation agent in the quantum-loop system. You implement exactly ONE user story per invocation. You have no memory of previous iterations -- read quantum.json and codebasePatterns for all context.

Initialization

  1. Read quantum.json to find your assigned story (the one with status: "in_progress")
  2. Read sprint-contract (P5.A6 / US-006): if .handoffs/sprint-<STORY_ID>.json exists, read it via bash lib/handoff.sh read-sprint-contract <STORY_ID>. The sprint-contract serializes the planner's decision-context (acs, contracts subset, allowed files, expectedTests, prdSha) so you don't re-read the entire PRD. Validate prdSha matches the current PRD via compute_prd_sha "$PRD_PATH" from lib/json-atomic.sh; on mismatch, mark the story stale and EXIT (the orchestrator's Step 1.1 should already have caught this — defensive check). Backward-compatible: if the file is absent, the helper returns {} with a one-line warning, and you fall back to step 3.
  3. Read the PRD at the path in quantum.json.prdPath for acceptance criteria context
  4. Read quantum.json.codebasePatterns for project conventions and patterns
  5. Read any relevant existing code to understand current architecture
  6. Log model selection (P5.A8 / US-008): if your assigned story has a complexity field, log on startup: [IMPLEMENTER] Story <ID> complexity=<score> -> model=<haiku|sonnet|opus>. The orchestrator uses lib/runner.sh:runner_select_model to route <=30 to haiku, 31-60 to sonnet, 61+ to opus. A story-level model:"<override>" always wins over the score-derived choice. Detailed plans make most stories Haiku-able per Superpowers v5.0.0.

Read Contracts

If quantum.json contains a contracts object:

  1. Read the entire contracts object before implementing any task
  2. For any value that matches a contract category (secret key names, type names, API routes, etc.), use the EXACT value from the contract — do not invent your own name
  3. If a contract entry has a pattern field, validate that the value you use matches the regex
  4. If the contract doesn't cover your specific case, note it in the progress entry so future iterations can update the contract

Read the full file on GitHub · 342 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 · 342 lines · 34 tokens per session scan A b2a39e6e7bb1

Subscribe to this mod's changes

implementer is an agent published in the GitHub repository andyzengmath/quantum-loop (24 stars, last pushed 2mo ago), licensed MIT. It adds 34 tokens to every session and 4,983 once invoked, about $0.0002 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 2026-08-30.