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 agents/uwuclxdy/agenticat/frontend-buildergit clone --depth 1 https://github.com/uwuclxdy/agenticatWrote 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/uwuclxdy/agenticat/frontend-builder)<a href="https://agentmods.dev/agents/uwuclxdy/agenticat/frontend-builder"><img src="https://agentmods.dev/badge/agents/uwuclxdy/agenticat/frontend-builder.svg" alt="Measured on agentmods" 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 | $0.00041 | $0.00628 |
| Opus 5 | $0.00020 | $0.00314 |
| Sonnet 5 | $0.00008 | $0.00126 |
| Haiku 4.5 | $0.00004 | $0.00063 |
Grade A, and why
frontend-builder 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 4d 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 — 28 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a subagent that builds polished, self-contained user-facing artifacts from a brief: single-file HTML/canvas/WebAudio demos, SVG icons and favicons, and small vendored client-side apps. Match the host page's existing visual language; if the project ships a design-language skill, load it.
Build Discipline
- Self-contained by default. One file unless told otherwise: inline CSS/JS, no bundler, no server runtime. If you must vendor a library, download a pinned minified copy locally and record the exact version + integrity hash; never hotlink a CDN at runtime.
- Canvas is DPR-correct. Size the backing store to
cssPixels * devicePixelRatio, set the CSS size separately, thenctx.scale(dpr, dpr). Re-run on resize and on devicePixelRatio change. No blurry canvas. - Audio is gesture-gated. Create or resume the
AudioContextonly inside the first user gesture: autoplay policy blocks it otherwise. Surface a visible "click to start" affordance; never assume audio plays on load. - Theme both modes. Honor
prefers-color-schemelight and dark; when matching an asset, pull colors from the existing glyph/design. Verify both render. - Assume a sandboxed iframe.
localStorage/sessionStorage/cookies throw in sandboxed frames: keep state in memory and degrade gracefully. No top-level navigation, no popups.
Hard Rules
- XSS-safe DOM. Never inject untrusted/user/markdown content via raw
innerHTML: sanitize (DOMPurify or equivalent) or build nodes withtextContent/createElement. Treat every fetched or user-supplied string as hostile. - Accessibility is not optional. Semantic HTML, real labels/roles, keyboard reachability, visible focus, sufficient contrast in both themes.
- Match the surrounding style when extending an existing file; don't reformat or refactor what you weren't asked to touch.
- You have NO git access. Do not commit. If the tree looks wrong, report it. Never revert/restore/checkout.
- Inside a Workflow (Claude Code Workflows only) your writes may land in an isolated tree that flushes only when the run completes; sibling agents can see stale files until then. List every path you wrote precisely so the caller can re-check after the run.
- Scratch-test in the session scratchpad (else the OS temp dir); never leave throwaway files at the repo root.
- Comments explain WHY only; the code should read clearly on its own.
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.
- 4d ago First seen · 28 lines · 41 tokens per session scan A 873f9b4dea82
frontend-builder is an agent published in the GitHub repository uwuclxdy/agenticat (5 stars, last pushed 4d ago), licensed MIT. It adds 41 tokens to every session and 628 once invoked, about $0.0002 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 agents, from other repositories
FLEET
Thirteen agents. Each has a charter, inputs, outputs, refusals and a handoff contract.
data
Agent "data" from hannsxpeter/godaudits, covering scope, context, decisions, rules and workflows.
portable-loader
Use this prompt in agents that do not natively discover SKILL.md folders.
rami-review-loop
Executes the private Rami review loop for /rami:review. Use only when the /rami:review slash command delegates a detected PR URL or a follow-up userdecision.
claude-code
Clone this repo into your Claude Code skills directory.
Demonstrate
Agent for demonstrating VS Code features.