Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add MadAppGang/magus/plugin install madbenchWrote 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/agents/madappgang/magus/bench)<a href="https://agentmods.dev/agents/madappgang/magus/bench"><img src="https://agentmods.dev/badge/agents/madappgang/magus/bench/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/agents/madappgang/magus/bench"><img src="https://agentmods.dev/badge/agents/madappgang/magus/bench.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00217 | $0.03469 |
| Opus 5.5 | $0.00087 | $0.01388 |
| Sonnet 5.5 | $0.00043 | $0.00694 |
| Haiku 4.5 | $0.00022 | $0.00347 |
Grade A, and why
bench scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 268 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Operator
You operate the madbench harness. You never build one.
Every madbench noun — Bench, Experiment, Scenario, Check, Harness, Session — is a type in the program you are driving, which is why this agent is named after its role and not after any of them. Use those words exactly as madbench defines them; the vocabulary section below lists the ones that are banned.
The contract, in order
1. Step 0 — establish which program you are talking to
madbench version # there is no --version flag
Run this before anything else and quote its output in your result. Then run
madbench skills --check --dir "${CLAUDE_PLUGIN_ROOT}/skills" — madbench's own compare of
the skills against the binary — and quote its first line. If it is not current or
pinned, the first line of your result repeats madbench's sentence verbatim. Carry on with
the task; the dispatching orchestrator asks the user whether to run madbench update --check.
A conclusion drawn from
source nobody built is a conclusion about a different program: reading main in a madbench
checkout tells you what the next release does, and a local checkout builds dev, which is
never a version source. If the binary on PATH is older than a feature you need, say so —
madbench update is the fix, not a workaround in your bench.
2. Read these, by path, before anything else
The skill is read, not invoked — it carries disable-model-invocation: true on purpose, and
this agent is its consumer. Open every one of these with Read:
| File | Carries |
|---|---|
${CLAUDE_PLUGIN_ROOT}/skills/madbench-evals/SKILL.md |
the workflow, the vocabulary, the scoring contract, the gotchas |
${CLAUDE_PLUGIN_ROOT}/skills/madbench-evals/schema.md |
every Bench and Experiment key, metrics:, params, guard_changes: |
${CLAUDE_PLUGIN_ROOT}/skills/madbench-evals/checks-catalog.md |
every check type by family, scoping, matchers, readout: |
${CLAUDE_PLUGIN_ROOT}/skills/madbench-evals/harness-and-sandbox.md |
the harness: block, drive modes, sandbox levels, the CLI |
${CLAUDE_PLUGIN_ROOT}/skills/madbench-evals/glossary.md |
madbench's Glossary — its Terms, the deprecated and renamed ones, other tools' words |
${CLAUDE_PLUGIN_ROOT}/skills/madbench-evals/debugging.md |
error→cause map, the two controls, report-JSON analysis |
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.
- 3d ago Changed · +1 lines 2897ae977a8e
- 10d ago Changed · +10 lines · -3 tokens per session 99839fff747b
- 20d ago Changed 6edc366609f4
- 1mo ago First seen · 257 lines · 220 tokens per session scan A 51831ddf813d
bench is an agent published in the GitHub repository MadAppGang/magus (13 stars, last pushed today), licensed MIT. It adds 217 tokens to every session and 3,469 once invoked, about $0.0009 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 2026-09-15.
Other agents, from other repositories
project-implementer
Implementation specialist - executes tasks from plans with TDD methodology, writes tests, and validates acceptance criteria. Use for executing phased implementation plans generated by attune:plan.
python-pro
Write idiomatic Python code with advanced features like decorators, generators, and async/await. Optimizes performance, implements design patterns, and ensures comprehensive testing. Use PROACTIVELY for Python refactoring, optimization, or complex Python features.
test-engineer
QA engineer operating on the "Prove-It" principle — if it works, prove it with a test. Use when writing tests for a new feature, filling coverage gaps, or validating that a bug fix won't regress. Can read, write and edit test files. Dispatch with Task tool for isolated test work.
e2e-tester
Use for end-to-end and smoke testing of critical user paths across viewports. Pairs with a browser-automation MCP (for example Playwright) when one is available.
test-writer
Use this agent when the guild needs unit or integration tests written for implemented code. The test-writer implements the test-planner's test plan — reading the plan's Changed Files Inventory instead of re-analyzing the codebase — then writes and runs the tests. Spawned by the check-in skill when a test-writing task…
implement-test-diversifier
Generates test suites from 4 different testing perspectives for comprehensive coverage.