Marin is an open-source research program, software platform, and community for developing foundation models such as large language models. Researchers use it for data preparation, tokenization, pretraining, posttraining, evaluation, and related experiments, including work on audio-text, DNA, and protein models. The catalogue entries are add-ons that support workflows around Marin.
Borrowing it
Nothing to install: this file belongs to marin-community/marin. 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/marin-community/marin/main/.agents/skills/marina-applet/SKILL.mdgit clone --depth 1 https://github.com/marin-community/marinWrote 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/marin-community/marin/marina-applet)<a href="https://agentmods.dev/skills/marin-community/marin/marina-applet"><img src="https://agentmods.dev/badge/skills/marin-community/marin/marina-applet/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/marin-community/marin/marina-applet"><img src="https://agentmods.dev/badge/skills/marin-community/marin/marina-applet.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 150 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 | $0.00072 | $0.01911 |
| Opus 5 | $0.00036 | $0.00955 |
| Sonnet 5 | $0.00014 | $0.00382 |
| Haiku 4.5 | $0.00007 | $0.00191 |
Grade A, and why
marina-applet 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 — 223 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Work with Marina applets
Read the Publishing an applet section of infra/marina/README.md before
changing an applet or publishing one. Use
infra/marina/examples/problem-set-applet/ as the minimal working package and
backend reference. Treat those files and marina --help as authoritative when
this skill disagrees with the checkout.
Choose the applet shape
Ensure the built package contains:
my-applet/
applet.toml
dist/index.html
Add browser assets below dist/. Use only relative asset, API, and query URLs
so both /a/<uuid>/ and /a/<uuid>/v/<revision>/ work.
The source directory may omit dist/index.html only when applet.toml has a
build_command that creates it before packaging.
Use plain HTML and JavaScript for a small single-view applet. Use Vue with Vite when the applet needs multiple views, reusable components, or substantial client state. For Vue applets:
- Set Vite's
baseto"./"so built asset URLs remain relative. - Use
createWebHashHistory()when the app has client-side routes. This avoids hard-coding an applet UUID or revision into the router base. - Use relative backend calls such as
fetch("api/results")andfetch("query", ...); do not start applet-owned URLs with/. - Bundle dependencies into
dist/. Marina's content security policy does not allow scripts loaded from a third-party CDN.
Keep the source index.html, package.json, the Vite configuration, and src/
beside applet.toml. They are local build inputs and are not uploaded. A
typical manifest uses build_command = "npm ci && npm run build". Marina runs
that command before every publish by default. Use --no-build only when the
existing dist/ has already been built and validated.
Use the applet's implicit Postgres schema for mutable or queryable data. Do not
create or attach a separate database. Load simple scalar tables with the CLI,
or add server/ when the applet needs parameterized business logic, custom
responses, or server-side integration.
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 Changed · +23 lines 1e26b24b39c1
- 4d ago Changed · +19 lines 6e7584bc31bb
- 5d ago First seen · 181 lines · 72 tokens per session scan A ca5e917e98f4
marina-applet is a skill published in the GitHub repository marin-community/marin (3,548 stars, last pushed today), licensed Apache-2.0. It adds 72 tokens to every session and 1,911 once invoked, about $0.0004 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 2026-09-05.
Other skills, from other repositories
copilotkit-upgrade
Use when migrating a CopilotKit v1 application to v2 -- updating package imports, replacing deprecated hooks and components, switching from GraphQL runtime to AG-UI protocol runtime, and resolving breaking API changes.
nextjs-pages-router
Set up tRPC in Next.js Pages Router with createNextApiHandler, createTRPCNext, withTRPC HOC, SSR via ssr option and ssrPrepass, SSG via createServerSideHelpers with getStaticProps, and server-side helpers for getServerSideProps prefetching.
with-tanstack-query
Compose Angular Query with signal-owned Table filtering, sorting, and pagination state using reactive query options, manual row-model boundaries, direct query data, server counts, and valid injection context.
auth-web-cloudbase
CloudBase Web Authentication Quick Guide for frontend integration after auth-tool has already been checked. Provides concise and practical Web authentication solutions with multiple login methods and complete user management.
service-digital-engagement-channel-configure
Configures and deploys enhanced chat Messaging Channels for Messaging for In-App and Web (MIAW). Use when the user needs to create, deploy, and activate a messaging channel configured with Omni-Channel Flow, Omni-Channel Queue, User, or Agentforce Service Agent routing. Generates MessagingChannel metadata, deploys it…
om-system-extension
Extend installed Open Mercato modules through UMES enrichers, interceptors, mutation guards, widgets, menus, entity extensions, events, component/page replacements, and overrides. Use for "extend core", "add field/column/action", "hide page", "intercept API", "UMES", or "rozszerz moduł".