Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/ninemindai/agentgemnpx agentmods add skills/ninemindai/agentgem/agentgem-miniappWrote 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/ninemindai/agentgem/agentgem-miniapp)<a href="https://agentmods.dev/skills/ninemindai/agentgem/agentgem-miniapp"><img src="https://agentmods.dev/badge/skills/ninemindai/agentgem/agentgem-miniapp/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/ninemindai/agentgem/agentgem-miniapp"><img src="https://agentmods.dev/badge/skills/ninemindai/agentgem/agentgem-miniapp.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 Excessive Agency · line 109 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00085 | $0.02949 |
| Opus 5.5 | $0.00034 | $0.01180 |
| Sonnet 5.5 | $0.00017 | $0.00590 |
| Haiku 4.5 | $0.00009 | $0.00295 |
Grade A, and why
agentgem-miniapp 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 2mo 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 — 205 lines — stays where its author put it; the contents beside it link to each section on GitHub.
agentgem-miniapp
Changing the miniapp platform, not authoring a miniapp? The normative spec is
docs/miniapps/spec.md— implement against its conformance checklist, and readdocs/miniapps/evolution.mdbefore re-deciding anything it records. If your change touches the authoring contract, editMINIAPP_BUILDER_BRIEFinpackages/play/src/builderBrief.ts— the body below is a byte-for-byte view of that constant (this file must end with it;src/play/__tests__/builderBrief.test.tsguards the mirror), and the Studio agent only ever sees the constant, so a SKILL.md-only edit changes nothing at runtime.
The file
You are editing one file: <name>.html — a single, self-contained HTML document. Never add a
second file. If the file has AGENTGEM:GAME-LOGIC start/end markers, keep your changes between
them.
It runs in a null-origin sandboxed iframe (sandbox="allow-scripts", no allow-same-origin)
under this Content-Security-Policy:
default-src 'none'; style-src 'unsafe-inline'; script-src 'unsafe-inline';
img-src data:; font-src data:; media-src data:;
So inline every byte of JS and CSS, use only data: URIs for images, fonts and media, and make no
network calls of any kind.
Theming. The host may supply CSS variables — --color-background-primary,
--color-background-secondary, --color-text-primary, --color-border-primary — and sets theme
("light" or "dark") on <html data-theme> automatically. Read them with a fallback, because most
hosts send none: background: var(--color-background-primary, #0d1117). On app.agentgem.ai, and any
other host that sends no variables, the fallback is what renders.
What Save enforces
Two gates run when the user saves. Both throw a message you will see in the studio.
The seal. Rejects, anywhere in executable code:
- an external
src=orhref=(anything other thandata:or#) - a bare module import (
import … from "…") - the words
fetch,XMLHttpRequest,WebSocket,EventSource,importScripts,navigator.sendBeacon
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.
- 2mo ago First seen · 205 lines · 85 tokens per session scan A 47d6d82c6863
agentgem-miniapp is a skill published in the GitHub repository ninemindai/agentgem (41 stars, last pushed 27d ago), licensed MIT. It adds 85 tokens to every session and 2,949 once invoked, about $0.0003 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.
agentgem-miniapp compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| agentgem-miniappninemindai/agentgem | Skill | 41 | 27d ago | MIT | A | 85 | Claude Code, Codex |
| demo-1-terminal-arcadedecodingai-magazine/building-a-coding-agent-from-scratch-course | Skill | 1.1k | 13d ago | Apache-2.0 | A | 46 | Claude Code, Codex |
| 3D模型与视频特效rojim666/SztuCode | Skill | 110 | 13d ago | MIT | A | 149 | Claude Code, Codex |
| vrchat-world-queriesggg123124/vrchat-assistant | Skill | 30 | 3d ago | MIT | A | 20 | Claude Code, Codex |
| test-designing-guidenowsprinting/unity-coding-skills | Skill | 20 | 8d ago | Unlicense | A | 52 | Claude Code |
| test-writing-guidenowsprinting/unity-coding-skills | Skill | 20 | 8d ago | Unlicense | A | 81 | Claude Code |
| run-testsnowsprinting/unity-coding-skills | Skill | 20 | 8d ago | Unlicense | A | 130 | Claude Code |