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.
npx agentmods add instructions/typedef-ai/ade-bench-plugin/claude-mdgit clone --depth 1 https://github.com/typedef-ai/ade-bench-pluginWrote 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.
[](https://agentmods.dev/instructions/typedef-ai/ade-bench-plugin/claude-md)<a href="https://agentmods.dev/instructions/typedef-ai/ade-bench-plugin/claude-md"><img src="https://agentmods.dev/badge/instructions/typedef-ai/ade-bench-plugin/claude-md.svg" alt="Measured on agentmods" height="20"></a>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 | $0.01477 | $0.01477 |
| Opus 5 | $0.00739 | $0.00739 |
| Sonnet 5 | $0.00295 | $0.00295 |
| Haiku 4.5 | $0.00148 | $0.00148 |
Grade A, and why
ade-bench-plugin CLAUDE.md 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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What this repository is
This is a Claude Code plugin (no application code, no build system, no tests). It ships configuration that is loaded by Claude Code in another user session — markdown skill/agent definitions plus knowledge/pattern data files. There is nothing to compile, run, or unit-test here. "Running" the plugin means: a user installs it, invokes a slash command (/ade-bench:plan-tasks or /ade-bench:create-task) inside their own dbt project, and Claude executes the instructions in skills/*/SKILL.md.
The plugin's purpose is to generate ADE-Bench benchmark tasks from a user's own dbt project. ADE-Bench tasks follow a setup/solve/verify pattern: a unified-diff patch introduces a bug into a working dbt model, the agent under test sees a symptom-only prompt, and pass/fail is determined by dbt tests + comparison of the output table against a seed CSV.
Repository layout and how the pieces interact
.claude-plugin/
plugin.json # Plugin manifest — name is "ade-bench" (the slash-command namespace)
marketplace.json # Local marketplace pointer — lets `claude plugin marketplace add <repo-path>` work
commands/ # Thin slash-command wrappers (so `/ade-bench:plan-tasks` etc. show in autocomplete)
skills/
plan-tasks/SKILL.md # Interactive pair-planning skill (the recommended path)
create-task/SKILL.md # Automated pipeline skill — invokes the discover-project agent
setup/SKILL.md # Installs ade-bench at ~/.ade-bench when missing; both skills above auto-invoke this on first use
agents/
discover-project.md # Read-only sub-agent used by create-task to scan a dbt project
knowledge/ # Prose .md files — used by plan-tasks as reasoning material
patterns/ # Structured .yaml files — used by create-task for token matching
example-output/ # Sample of what a generated task directory looks like
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.
- 5d ago First seen · 60 lines · 1,477 tokens per session scan A 657d0e24d784
ade-bench-plugin CLAUDE.md is an instructions file published in the GitHub repository typedef-ai/ade-bench-plugin (3 stars, last pushed 3mo ago), licensed MIT. It adds 1,477 tokens to every session, about $0.0074 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.
Other instructions, from other repositories
docglow CLAUDE.md
Instructions for docglow/docglow, covering claude.md — docglow, project, tech stack, build & test and python.
dex copilot-instructions.md
Copilot instructions for exmergo/dex: Follow AGENTS.md at the repository root.
dbt-doctor AGENTS.md
Instructions for northgraindata/dbt-doctor, covering dbt-doctor, stack, commands and conventions.
dex AGENTS.md
AGENTS.md instructions for exmergo/dex, covering dex: driving the engine from any agent, the loop: explore, transform, maintain (etm), the command contract, or, with the pinned wrapper a skill ships and the envelope.
erd-studio CLAUDE.md
Instructions for liam-machine/erd-studio, covering claude.md, naming, directory structure, internal identifiers and build & test commands.
river-review GEMINI.md
Gemini CLI instructions for s977043/river-review, covering gemini instructions (river-review), 必須チェック, gemini-specific tips, prompt assembly and development tips.