A guided pair-programming workflow for test-driven development, or TDD: writing a failing test, making it pass, then improving the code. It runs these red, green, and refactor cycles in batches and keeps progress in a project state file.

Use it to implement a finalized specification through staged test cycles, with interactive review or an autonomous review process at each phase boundary.

It gives implementation a repeatable loop and checks each phase against a specification. Reviews can be done by a person or by an automated reviewer, depending on the mode.

tdd-pair — Skill for Claude Code by mandelbro/agent-workflow-kit

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.

Any agent

npx skills add mandelbro/agent-workflow-kit --skill tdd-pair

Clone the repo

git clone --depth 1 https://github.com/mandelbro/agent-workflow-kit

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/skills/mandelbro/agent-workflow-kit/tdd-pair/github.svg)](https://agentmods.dev/skills/mandelbro/agent-workflow-kit/tdd-pair)

Your own site

<a href="https://agentmods.dev/skills/mandelbro/agent-workflow-kit/tdd-pair"><img src="https://agentmods.dev/badge/skills/mandelbro/agent-workflow-kit/tdd-pair/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/skills/mandelbro/agent-workflow-kit/tdd-pair"><img src="https://agentmods.dev/badge/skills/mandelbro/agent-workflow-kit/tdd-pair.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 113 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 7,841 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. 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	$0.00113	$0.07841
Opus 5.5	$0.00045	$0.03136
Sonnet 5.5	$0.00023	$0.01568
Haiku 4.5	$0.00011	$0.00784

Measured 9d ago against content hash 72f017d3913a, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

tdd-pair 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 9d 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.

src/claude_workflow_kit/assets/skills/tdd-pair/SKILL.md · 413 lines

How it starts

The opening of the file, as written. The whole thing — 413 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/tdd-pair — TDD Loop Pairing Session

You are operating as Ada (the pair-programming persona; use the Ada agent definition if one is installed, otherwise follow this file alone) in TDD Loop Pairing mode. This skill orchestrates the inner loop of the TDD-with-AI Workflow's step 6 ("Pipelined implementation"): turning a finalized spec into batched red/green/refactor cycles, with you in the typing role and an auditor in the review role at every phase boundary.

Two modes, one engine. The red→green→refactor state machine is identical regardless of who audits; only the gate at each phase boundary differs:

Interactive mode (start) — the human is the auditor. Each phase boundary is an AskUserQuestion. Use when you want input into the process.
Autonomous mode (auto) — the Zod reviewer agent is the auditor. Each phase boundary is a Zod review against the spec, with bounded retry and escalation to the human on impasse. Use for hands-off, contract-driven implementation (e.g. the Task Workflow System's implementation step).

See the Gate Protocol section for how the two gates resolve. Everything else — batching, state persistence, the phase flows — is shared.

Headless variant. auto --headless (and resume --headless) is autonomous mode for unattended runs such as claude -p. It never calls AskUserQuestion or PushNotification; wherever another flow says "STOP and surface", it ends the run with a JSON outcome instead. See Headless Mode.

Subcommand input: $ARGUMENTS

Parse the first whitespace-separated token of $ARGUMENTS to determine the subcommand. Remaining tokens are subcommand arguments. Supported subcommands: start, auto, resume, status, end. auto and resume accept a --headless flag anywhere in the arguments; strip it before reading the spec path. If no subcommand or an unknown one is given, print the help block at the bottom of this file and exit.

State File

Location: <cwd>/.claude/tdd-state.json (per-project, survives reboots, gitignore-able)

Read the full file on GitHub · 413 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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.

9d ago First seen · 413 lines · 113 tokens per session scan A 72f017d3913a

Subscribe to this mod's changes

tdd-pair is a skill published in the GitHub repository mandelbro/agent-workflow-kit (2 stars, last pushed 9d ago), licensed Apache-2.0. It adds 113 tokens to every session and 7,841 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-10-31.

TDD Testing

tdd-pair next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
tdd-pairmandelbro/agent-workflow-kit	Skill	2	9d ago	Apache-2.0	A	113	Claude Code
engram-testing-coverageGentleman-Programming/engram	Skill	7.9k	2d ago	MIT	A	25	Claude Code, Codex
principle-test-behavior-not-implementationmichael-denyer/pstack-claude	Skill	4.1k	yesterday	MIT	A	50	Claude Code
tddcodewithmukesh/dotnet-claude-kit	Skill	808	5d ago	MIT	A	81	Claude Code
nw-ddd-eventsourcingnWave-ai/nWave	Skill	633	2d ago	MIT	A	38	Claude Code
nw-fp-clojurenWave-ai/nWave	Skill	633	2d ago	MIT	A	25	Claude Code
nw-leanux-methodologynWave-ai/nWave	Skill	633	2d ago	MIT	A	37	Claude Code