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 agentmods add skills/lengros/bennu/prototypenpx skills add Lengros/bennu --skill prototypegit clone --depth 1 https://github.com/Lengros/bennuWhat 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 | $0.00251 | $0.04141 |
| Opus 5 | $0.00125 | $0.02070 |
| Sonnet 5 | $0.00050 | $0.00828 |
| Haiku 4.5 | $0.00025 | $0.00414 |
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 2d 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 — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/prototype — wire screens into one clickable flow, render and CLICK it
This skill makes a navigable flow you can click through, not a pretty dead-end frame.
It owns the gap /design leaves: /design hand-codes ONE static screen and renders it to a
PNG; /diagram draws how a system is wired. Neither lets a PM open a file and walk the
path — tap a button, land on the next screen, watch a value update. That clickable-prototype
gap is real and Bennu had not closed it before this skill.
Core principle: A clickable prototype validates a FLOW, not a look — so the thing you
must lock before any HTML is the SCREEN MAP and the NAVIGATION GRAPH, not the art direction.
Get the inventory of screens and, for every hotspot, which screen it leads to and what state
it changes — then the same self-contained-HTML discipline /design already uses produces a
working flow instead of pretty cul-de-sacs.
This is the interaction counterpart to /design (identity) and /diagram
(structure). It reuses their renderer and skeptic loop; the only new thing it teaches is
multi-screen wiring inside one file.
The one thing that makes this not /design
/design's hard gate is "lock the art direction before HTML." /prototype's hard gate is:
lock the SCREEN MAP + NAVIGATION GRAPH before HTML. Skip it and you get beautiful screens
that go nowhere — the exact failure this skill exists to prevent. Fidelity is a dial, not
the point: default to the lowest fidelity that proves the flow.
How to run
Stages in order. Hard gate: no HTML until the screen map + nav graph are locked (Stage 2).
1. Brief (intake — keep it to these lines)
Settle before building. Ask only what's missing from context; don't interrogate.
- flow — the one user journey to validate (e.g. "first-run onboarding", "add a payee and pay")
- key path — the single end-to-end sequence the prototype MUST demo, screen by screen. This is the spine everything else hangs off; without it the prototype sprawls.
- actors & decision — who clicks through it, and what PM/design decision it must unblock
- fidelity — lo-fi wireframe (greyscale boxes, prove the flow) / mid (real layout,
placeholder polish) / hi-fi (looks shippable). Default lo-fi; raise only with a reason.
Hi-fi visual identity is
/design's job, not this skill's — borrow tokens, don't compete. - simulated state — what must change on interaction (a form value echoed back, a toggle, a list that grows, a "logged-in" flag that persists across a reload)
- working path — settled now: default
.scratch/prototype/<slug>.htmlfor exploration, or the target project's home for real work. Every later command uses this one path.
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.
- 2d ago First seen · 263 lines · 0 tokens per session scan A 6791ed441d9e
prototype is a skill published in the GitHub repository Lengros/bennu (2 stars, last pushed 1mo ago), licensed MIT. It adds 251 tokens to every session and 4,141 once invoked, about $0.0013 per session on Opus 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 2026-08-31.
Other skills, from other repositories
frontend-design
Produce intentional, responsive, accessible UI work and perform visual QA instead of generic component assembly.
tutti-ui-system
Use when working with @tutti-os/ui-system components, replacing local UI with shared components, querying component ids or metadata, promoting UI into shared base or business components, or maintaining UI-system storyboard inventory.
frontend-design
Use this as the design-lead layer for VideoStudio COMPOSE work. It shapes project/composition/composition-manifest.json::artdirection and the model-authored HTML/SVG motion graphics in index.html.
ui-artifact-workspace
Use this skill for every standalone UIDesigner artifact, regardless of final format, and for every follow-up that revises a previously created artifact. It is the continuity and packaging contract around the visual skills; it does not decide the visual direction itself.
ui-design-executor
Use this as UIDesigner's default execution skill. It contains the baseline design, accessibility, HTML, runtime-safety, responsive, taste, and verification rules needed by ordinary UI work. Do not load the separate system, control, taste, color, renderer, or craft skills merely to repeat these baseline rules.
ui-craft-checks
Use this comprehensive gate for formal review, QA, launch handoff, exact-fidelity inspection, high-risk complex UI, or when the fast gate exposes a deeper craft problem. Ordinary single-screen work uses ui-design-executor and its bundled validator without loading this full matrix. Pair standalone artifacts with…