Borrowing it
Nothing to install: this file belongs to WolframResearch/AgentTools. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/WolframResearch/AgentTools/main/.claude/skills/verify/SKILL.mdgit clone --depth 1 https://github.com/WolframResearch/AgentToolsWrote 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/wolframresearch/agenttools/verify)<a href="https://agentmods.dev/skills/wolframresearch/agenttools/verify"><img src="https://agentmods.dev/badge/skills/wolframresearch/agenttools/verify/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/wolframresearch/agenttools/verify"><img src="https://agentmods.dev/badge/skills/wolframresearch/agenttools/verify.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium MCP Rug Pull · line 19 npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.Fix: Pin the version: npx @scope/[email protected]
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 | $0.00030 | $0.00761 |
| Opus 5.5 | $0.00012 | $0.00304 |
| Sonnet 5.5 | $0.00006 | $0.00152 |
| Haiku 4.5 | $0.00003 | $0.00076 |
Grade A, and why
verify scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 23 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Verifying AgentTools changes at runtime
Wolfram Language side (tools, server, UI results)
- Always set
WOLFRAMSCRIPT_KERNELPATH="C:/Program Files/Wolfram Research/Wolfram/15.0/wolfram.exe"(or newer). Plainwolframscriptmay pick an older kernel; AgentTools requires 15.0+, and under an older kernelPacletDirectoryLoad+Get["Wolfram`AgentTools`"]silently falls back to the installed Repository paclet — you exercise the wrong code. Confirm withFindFile["Wolfram`AgentTools`"]→ must resolve under the dev directory. - Load the dev paclet:
PacletDirectoryLoad["H:\\Documents\\AgentTools"]; Get["Wolfram`AgentTools`"]. - To drive a tool exactly as the MCP server dispatches it, call its registered
"Function"(see$defaultMCPTools[name]inKernel/Tools/*.wl) with the parsed-args association, e.g.Wolfram`AgentTools`Tools`WolframLanguageEvaluator`Private`evaluateWolframLanguage[<|"code" -> ...|>]. - To trigger the MCP Apps UI path (cloud-deployed notebook +
_meta.notebookUrl), first setWolfram`AgentTools`Common`$clientSupportsUI = Trueand be cloud-connected. Export the result withDeveloper`WriteRawJSONFile— it is the exact content/_meta payload a host forwards to a viewer. - For an A/B against pre-change behavior,
git worktree add <scratch>/head-copy HEADand pointPacletDirectoryLoadat the worktree in a second wolframscript run. Use a different input (the deploy target is content-hashed, same input overwrites the same cloud object). Remove the worktree afterwards.
MCP Apps viewers (Assets/Apps/*.html)
- All viewers speak the same iframe protocol: they send
ui/initialize(reply with{protocolVersion, hostInfo, hostContext}), then notifyui/notifications/initialized; after that, post{jsonrpc:"2.0", method:"ui/notifications/tool-result", params:{content, _meta}}into the iframe. - Harness that works: a
host.htmlthat iframes the viewer and plays the host role over postMessage + a no-dep node static server + Playwright (npm i playwrightlocally; chromium vianpx playwright install chromium). SetcolorScheme: "dark"on the browser context to reproduce dark-host bugs; the embed needs real network (unpkg CDN + wolframcloud.com) and takes 10–60 s. - The notebook embedder (
useShadowDOM: true) attaches an open shadow root to a wrapper<div>it creates inside the container, not to the container itself. Wait for embed completion by recursively searching open shadow roots for expected notebook text; read computed styles withgetComputedStyleon shadow-tree elements. - Playwright frame-picking gotcha: the host page URL carries the viewer filename in its query string, so match
f !== page.mainFrame()when locating the viewer frame. - Verified evidence to capture: computed
colorof shadow text leaves (dark-theme leak shows asrgb(224, 224, 224)), computedfont-family, and full-page screenshots.
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.
- 1mo ago First seen · 23 lines · 30 tokens per session scan A fc7f3dd90cf6
verify is a skill published in the GitHub repository WolframResearch/AgentTools (96 stars, last pushed yesterday), licensed MIT. It adds 30 tokens to every session and 761 once invoked, about $0.0001 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 2026-08-30.
Other skills, from other repositories
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.
verify-omnigent
Drive Omnigent the way a user does and prove a behavior with recorded evidence, using an isolated server, runner, and mock model plus a feature map of every user entry point. Load before reproducing a user-facing bug, before claiming a fix works, or when reviewing whether a change covered every surface (session vs.…
test-master
Generates test files, creates mocking strategies, analyzes code coverage, designs test architectures, and produces test plans and defect reports across functional, performance, and security testing disciplines. Use when writing unit tests, integration tests, or E2E tests; creating test strategies or automation…
dogfood
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.
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.
critique
Adversarial design critique of the current work — render it, look at it, and argue for rejection. Run after the gates are green, never instead of them.