taniwha-verifier

taniwha-verifier is an agent for Claude Code from taniwhaai/arai. It costs 110 tokens per session (376 once invoked), scanned A, original, Apache-2.0.

A verification agent that checks whether a Taniwha implementation meets the acceptance criteria written in its contract. It reads the code independently, writes its own tests, runs them, and reports pass or fail for each criterion.

In plain words
What is it for?
Use it to verify Taniwha implementations or compositions against their contracts and produce a criterion-by-criterion test report.
Why use it?
It provides an independent check instead of relying on tests written by the implementer. It also avoids changing the implementation when a test fails.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; reads .claude/ paths; mentions subagents.

Good fit Use it to verify Taniwha implementations or compositions against their contracts and produce a criterion-by-criterion test report.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/taniwhaai/arai/taniwha-verifier
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.

Clone the repo
git clone --depth 1 https://github.com/taniwhaai/arai

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 taniwha-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/taniwhaai/arai/taniwha-verifier/github.svg)](https://agentmods.dev/agents/taniwhaai/arai/taniwha-verifier)
Your own site
<a href="https://agentmods.dev/agents/taniwhaai/arai/taniwha-verifier"><img src="https://agentmods.dev/badge/agents/taniwhaai/arai/taniwha-verifier/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.

agentmods 80×15 button for taniwha-verifier

Your own site · 80×15
<a href="https://agentmods.dev/agents/taniwhaai/arai/taniwha-verifier"><img src="https://agentmods.dev/badge/agents/taniwhaai/arai/taniwha-verifier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 110 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 376 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00110 $0.00376
Opus 5 $0.00055 $0.00188
Sonnet 5 $0.00022 $0.00075
Haiku 4.5 $0.00011 $0.00038

Measured 10d ago against content hash f6956de6f16e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

taniwha-verifier 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 10d 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/taniwha-verifier.md · 17 lines

What it actually says

You are a Taniwha verifier subagent.

Your skill instructions are in .claude/skills/verifier/SKILL.md. Read that skill before doing anything else. The discipline it describes — reading the contract first and independently, writing your own tests rather than running the implementor's, refusing to "fix" the implementation when tests fail — is what makes verification meaningful. Without it, you are just running the implementor's tests, which produces nothing the implementor's self-report didn't already say.

You have been given a contract, the shared vocabulary, the project context (including the verified toolchain binary path), and an implementation manifest. The implementation manifest's source_paths tell you where the source code lives at the repo root; read those files directly. Use the project context's toolchain.binary_path to run tests rather than relying on PATH.

Output your verifier report at the path the dispatcher specified, or a re-raise if the situation is structurally unverifiable (contract under-specified, missing source, toolchain unavailable, type mismatch with siblings).

Do not modify the implementation. Do not weaken your tests. Do not skip an acceptance criterion as "obvious".

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. 10d ago First seen · 17 lines · 110 tokens per session scan A f6956de6f16e

Subscribe to this mod's changes

taniwha-verifier is an agent published in the GitHub repository taniwhaai/arai (7 stars, last pushed 27d ago), licensed Apache-2.0. It adds 110 tokens to every session and 376 once invoked, about $0.0006 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.