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.
git clone --depth 1 https://github.com/MadAppGang/magusWrote 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/plugins/madappgang/magus/madbench)<a href="https://agentmods.dev/plugins/madappgang/magus/madbench"><img src="https://agentmods.dev/badge/plugins/madappgang/magus/madbench/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.
<a href="https://agentmods.dev/plugins/madappgang/magus/madbench"><img src="https://agentmods.dev/badge/plugins/madappgang/magus/madbench.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
madbench 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.
How it starts
The opening of the file, as written. The whole thing — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
{
"name": "madbench",
"version": "0.11.0",
"description": "Toolkit for madbench, MadAppGang's Go harness for benchmarking agentic coding tools. Carries madbench's own skills with /madbench:madbench as the entry point, checks them against the installed madbench, and runs benches in a visible pane through the madbench:bench agent.",
"author": {
"name": "MadAppGang",
"email": "[email protected]"
},
"license": "MIT",
"homepage": "https://github.com/MadAppGang/magus",
"tags": [
"madbench",
"evals",
"benchmarking",
"agent-testing",
"assertions",
"llm-judge",
"go"
],
"dependencies": [
{
"name": "terminal",
"version": "^5.0"
}
],
"skills": "./skills/",
"conventions": {
"claudemd": {
"section": "madbench",
"template": "./templates/claude-md-conventions.md",
"version": "1.2.0"
}
},
"compatibility": {
"claudeCode": ">=0.1.0"
},
"requires": {
"schema": 1,
"bin": [
{
"name": "bun",
"check": [
"bun",
"--version"
],
"install": [
{
"via": "bootstrap",
"tool": "bun"
}
]
},
{
"name": "madbench",
"check": [
"madbench",
"version"
],
"install": [
{
"via": "brew",
"os": [
"darwin"
],
"cask": "madappgang/tap/madbench"
},
{
"via": "release",
"repo": "MadAppGang/madbench-releases",
"tag": "v0.46.0",
"asset": "madbench_{os}_{arch}.tar.gz",
"binary": "madbench",
"sha256": {
"darwin_amd64": "02f8f6a09d66268190e265b516de0f1e6635116506234ab04638969d1d6ad244",
"darwin_arm64": "fcbc2e979f3fc3afdb15d804544c14a4d7442cbd0b2112fdb61221ec61ab0c4f",
"linux_amd64": "0c4b25dac7ea4ee9b18cda8000ab7bb94945a29a2What it installs
The manifest is a name and a version. 3 skills, 1 command, 1 agent travel with it, and installing the plugin installs all of them — 411 tokens a session between them. Each is measured on its own page, and each can be installed alone.
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.
- yesterday Changed 8208ec214ee7
- 3d ago Changed 5bb4167f13ff
- 9d ago Changed 33dd5412dec1
- 10d ago Changed · +76 lines a3df7ecc1360
- 20d ago Changed 4297d087fac1
- 22d ago Changed dff36d74c424
- 25d ago Changed f5939023297f
- 1mo ago Changed ed2510ed7b40
- 1mo ago Changed · +13 lines 13ee887cb562
- 1mo ago First seen · 25 lines scan A f9dc1796aa66
madbench is a plugin published in the GitHub repository MadAppGang/magus (13 stars, last pushed today), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 2026-09-04.
Other plugins, from other repositories
load-tester
Load and stress testing for APIs and web services.
kit
Spec-driven Claude Code workflow with verification pipeline. Hooks, slash commands, and verification subagents.
headless
Headless browser automation for site comparison, E2E testing, and anti-bot-aware web research.
pict-test-designer
Design comprehensive test cases using PICT (Pairwise Independent Combinatorial Testing) for any piece of requirements or code. Analyzes inputs, generates PICT models with parameters, values, and constraints for valid scenarios using pairwise testing. Outputs the PICT model, markdown table of test cases, and expected…
local-test-plugin
Symlink-based local plugin testing for development workflow.
claude-md-management
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.