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 Mathews-Tom/armory --skill prototypegit clone --depth 1 https://github.com/Mathews-Tom/armoryWrote 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/mathews-tom/armory/prototype)<a href="https://agentmods.dev/skills/mathews-tom/armory/prototype"><img src="https://agentmods.dev/badge/skills/mathews-tom/armory/prototype/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/mathews-tom/armory/prototype"><img src="https://agentmods.dev/badge/skills/mathews-tom/armory/prototype.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 | $0.00174 | $0.02546 |
| Opus 5.5 | $0.00070 | $0.01018 |
| Sonnet 5.5 | $0.00035 | $0.00509 |
| Haiku 4.5 | $0.00017 | $0.00255 |
Grade A, and why
prototype 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Prototype
A prototype is throwaway code that answers one question. The question decides the shape, the artifact lives in scratch space, and the answer comes from observing the artifact, not from reading its source. The verdict feeds separate implementation work; the prototype itself never becomes production code.
When to Use
| Situation | Package |
|---|---|
| "Does this state model / rule / transition logic hold up?" | prototype (logic branch) |
| "What should this screen look like? Show me a few structurally different options" | prototype (UI branch) |
Working a decision-map:prototype ticket (claim, ticket evidence, frontier) |
decision-map, which builds the artifact with this package |
| A finished, polished HTML deliverable: dashboard, infographic, interactive diagram | static-web-artifacts-builder |
| Audit or redesign an existing page against UX criteria | ux-expert |
Prerequisites and Authority
These boundaries hold in every branch:
- Scratch only. The artifact lives in a scratch location discovered from the repository (step 3), never in production paths or tracked implementation files.
- No production-route edits. To show a variant in its real surroundings, build a scratch copy or read-only reproduction of the host page and record the realism you gave up. Never modify a live route, layout or shared component to host the experiment.
- Synthetic, in-memory data. Representative, non-sensitive data held in memory. No production database, credentials, persistence, live mutations or dependency installation. If the question is about persistence itself, use explicitly disposable scratch storage, say so, and never connect production storage; if the question cannot be answered under these rules, stop and report the missing prerequisite.
- No branch or commit by default. Commit the prototype only if the user explicitly asks, and then only to a new throwaway branch (see Output).
- Never lift prototype logic into production. No extraction, winner promotion or automatic cutover. The verdict is input to separately authorized implementation work.
- Browser before verdict. An empirical verdict requires observing the actual artifact in a browser (any authorized browser automation or a browser the user drives and reports back from). Anything not exercised is labelled not observed.
- Skip the polish. Only the structure and error handling needed to run the experiment. No new framework setup, reusable architecture or test suite for the prototype; a UI scratch copy may use the project's existing framework when
references/ui.mdreuses its runner.
What ships with it
6 files 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.
- yesterday First seen · 145 lines · 174 tokens per session scan A 415c66a6fdf2
prototype is a skill published in the GitHub repository Mathews-Tom/armory (337 stars, last pushed yesterday), licensed MIT. It adds 174 tokens to every session and 2,546 once invoked, about $0.0007 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 2026-10-27.
Other skills, from other repositories
make-landing
A design workflow for creating several visual versions of a landing page, hero area, cover, or slide surface. Each version is previewed, shown to a person for selection, and the chosen direction is then applied to the main project.
art-director
Orchestrate iterative visual style searches with branch prompts, decision graphs, feedback loops, and final direction selection.
html-draft
Use when user wants a standalone HTML diagram in flat engineering blueprint style — architecture diagrams, system flows, technical spec sheets, component maps. Generates one HTML file using Tailwind v4 (browser CDN) for layout and D3 v7 (CDN) for SVG diagrams. User-invoked only — do NOT auto-trigger. Triggers on…
make-3d
A 3D design workflow that creates several versions of a model in separate copies of a scene. The versions are shown using the same example data so a person can choose one before it replaces the main model.
idea-superpowers-suite
Use when running the full idea workflow: capture a rough idea, expand it into design/UI/implementation docs, research similar products, and generate build-ready Markdown artifacts.
idea-to-design-doc
Use when turning a rough idea into a focused product/design Markdown doc through guided questions, without moving into implementation too early.