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 skills add MadAppGang/magus --skill madbenchgit 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/skills/madappgang/magus/madbench)<a href="https://agentmods.dev/skills/madappgang/magus/madbench"><img src="https://agentmods.dev/badge/skills/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/skills/madappgang/magus/madbench"><img src="https://agentmods.dev/badge/skills/madappgang/magus/madbench.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 | $0.00052 | $0.01964 |
| Opus 5.5 | $0.00021 | $0.00786 |
| Sonnet 5.5 | $0.00010 | $0.00393 |
| Haiku 4.5 | $0.00005 | $0.00196 |
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 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.
How it starts
The opening of the file, as written. The whole thing — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/madbench — create and run madbench benches
The one entry point for madbench work. It reads the vocabulary, checks versions, reads the request, loads the reference, and does the work — in that order.
Request: $ARGUMENTS
Step 0 — the Glossary and versions, before anything else
Glossary first. Read ../madbench-evals/glossary.md, relative to this skill's
directory. Write only its words for madbench concepts; translate the user's words, never echo
them.
Then the versions. First run madbench skills --check and quote its output. It answers
the whole question: its first line is skills <W> · madbench <I> · <state>, the state one of
current, stale, pinned, disagree or unmeasured. Exit 0 means current or pinned,
1 means stale or disagreeing stamps, and 3 means there is nothing to compare. When it
answers, the version check is done; go to step 4.
Only when madbench skills does not exist (a madbench older than 0.38.0 answers
unknown command or tries to load skills as a bench), compare by hand:
- Read
MADBENCH_VERSIONin this skill's directory. It holds the madbench release these skills were written for, one line,x.y.z.madbench skillswrote it; never write one yourself. - Run
madbench version. It printsmadbench x.y.z. There is no--versionflag. - Compare the two numbers part by part (0.9.0 is lower than 0.33.1). With X the installed
madbench and Y the value of
MADBENCH_VERSION, say exactly one of these:
| When | Say, verbatim with the numbers filled in |
|---|---|
| Y equals X | madbench X · skills for X |
MADBENCH_PINNED exists beside MADBENCH_VERSION and Y differs from X |
madbench X · skills pinned to Y — a pin is deliberate, so it is not stale |
| Y is LOWER than X — the skills are behind | "You're using madbench X, but your madbench skills were written for Y. Update the skills: run madbench skills in this project, or update the plugin that installed them." |
| Y is HIGHER than X — madbench is behind | "Your madbench skills were written for Y, but madbench X is installed. Update madbench: madbench update." |
madbench dev, no madbench on PATH, or no MADBENCH_VERSION |
say which one, and that nothing was compared |
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 10d ago First seen · 133 lines · 52 tokens per session scan A fffad978958a
madbench is a skill published in the GitHub repository MadAppGang/magus (13 stars, last pushed today), licensed MIT. It adds 52 tokens to every session and 1,964 once invoked, about $0.0002 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 2026-10-07.
Other skills, from other repositories
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
eval-dataset-design
A guide to creating reliable test sets for AI agents. Each evaluation task includes the request, the simulated user's behaviour, a reset starting state, and a success check that can be independently verified.
tika-eval-compare
Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.
atmos-validation
Validate Atmos projects, components, arbitrary JSON Schema inputs, EditorConfig, and GitHub Actions; use affected-file selection and native CI annotations.