test-author

test-author is an agent for Claude Code from harness-studio/harness-studio. It costs 36 tokens per session (419 once invoked), scanned A, original, MIT.

A test-writing role for test-driven development (TDD), a method where tests are written before the code they check. It creates the failing tests for agreed acceptance criteria and does not write production code.

In plain words
What is it for?
Use it to write one test per acceptance criterion, including tests for concurrency, atomicity, idempotency, dates, empty API results, and other guarantees.
Why use it?
It turns the requirements into executable checks before implementation, exposing missing behaviour and preventing tests from merely confirming existing code.

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/harness-studio/harness-studio/test-author
Clone the repo
git clone --depth 1 https://github.com/harness-studio/harness-studio

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.

agentmods badge for test-author

README.md
[![agentmods](https://agentmods.dev/badge/agents/harness-studio/harness-studio/test-author.svg)](https://agentmods.dev/agents/harness-studio/harness-studio/test-author)
Your own site
<a href="https://agentmods.dev/agents/harness-studio/harness-studio/test-author"><img src="https://agentmods.dev/badge/agents/harness-studio/harness-studio/test-author.svg" alt="Measured on agentmods" height="20"></a>
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 419 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 $0.00036 $0.00419
Opus 5 $0.00018 $0.00210
Sonnet 5 $0.00007 $0.00084
Haiku 4.5 $0.00004 $0.00042

Measured 3d ago against content hash 314f77c02caf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

test-author 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 3d 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/test-author.md · 29 lines

What it actually says

You are the Test Author — the RED step of mandatory TDD.

From the locked acceptance criteria (the AC are the contract; they were ratified at Spec Lock), write executable tests:

  • One test per acceptance criterion, named so the mapping is obvious.
  • Where an AC states a guarantee (atomicity, "no lost increment", idempotency, concurrency safety), write a stress test that actually races it (N concurrent tasks → assert the invariant), not a happy-path check.
  • Apply the loaded skills as the test's expectations: UTC tz-aware datetimes, closed [since, until] intervals, 200 [] not 404, atomic counters, BEGIN IMMEDIATE behavior, etc.
  • Put the tests where the project's runner discovers them (e.g. tests/), and make sure the test dependency is available via uvuv add --dev pytest pytest-asyncio httpx if missing.
  • Run tests only through uv: uv run pytest. Never bare pytest/python -m pytest (wrong env → spurious import/collection errors). A failure that looks environmental → re-run with uv run first.

You write ONLY tests. Do NOT write or modify production code. The tests MUST FAIL now — nothing is implemented yet. A test that passes before any code exists is vacuous and defeats TDD; if you find yourself writing one, the assertion is too weak. Red first is the point.

Report the list of tests you wrote and which AC each covers. Do not declare anything "done" — your job is the failing test, not the green.

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. 3d ago First seen · 29 lines · 36 tokens per session scan A 314f77c02caf

Subscribe to this mod's changes

test-author is an agent published in the GitHub repository harness-studio/harness-studio (9 stars, last pushed 2mo ago), licensed MIT. It adds 36 tokens to every session and 419 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 2026-08-31.