플러그인 순위

Terminal Commands

Run commands on your machines

Web & Mobile App Builder

Build web and mobile apps

ShardX Proxy Advisor

Choose and fix proxy setups

Dashboard Icons

Homarr Labs

Find and preview icons for services, applications, and self-hosted dashboards. Match several service names, search the catalog, browse icons and categories, and retrieve CDN URLs, available variants, SVG markup, and license information. ChatGPT displays icon thumbnails alongside links. The public connection provides six read-only tools and requires no account. It cannot submit icons, manage accounts, or change the catalog.

TokenOS Gas Tracker

Live fees on five chains

ohmyho.st

Build & host websites and apps

AdSimple Consent Manager

AdSimple GmbH

AdSimple Consent Manager helps website owners understand their cookie consent setup. Connect your AdSimple account to ask questions about your websites using the data already available in your Consent Manager. You can list your domains, inspect consent settings, review detected cookies and external services, check the latest website scan and banner installation status, and retrieve the banner embed code. The connector is read-only. It cannot change banner settings, start scans, modify cookie categories, or manage billing. It reports recorded scan results and installation status, not a new live audit or a guarantee of legal compliance. Account content, including cookie descriptions, may be in German. Sign in with your adsimple.at account and review the domains shown before allowing access. You can revoke access in the Consent Manager under "Verbundene KI-Assistenten". Data retrieved through the connector is shared with OpenAI to answer your request. See the privacy policy for details.

BCD — Bridge Control Delegate

HEESEONG KIM

BCD securely connects ChatGPT and Codex to computers you explicitly authorize. Inspect and edit files, run shell commands, and manage durable processes across paired devices. New computers start read-only, with write and process access enabled only when you approve it.

Ample Computer

Ample Computer

Describe an app or share your code, and build and host it with Ample Computer. Get a live HTTPS URL without configuring servers or a cloud console. Create a Secret Santa app that saves participants and previews assignment emails, a booking form that stores reservations in a database, or a gallery that accepts and serves uploaded photos. Ample detects the app stack, builds it, connects managed databases and file storage, and keeps releases so you can update the same app. Inspect apps, read build and runtime logs, and fix failed deployments from the conversation. Sign in with your Ample Computer account in the browser. Preview app emails before sending. Image handling means uploading, storing and serving files, not generating images.

AgentAvow

Kenne Ives

AgentAvow tells you whether a tool is safe before your agent connects to it. Give it an MCP server, a package from npm, PyPI, crates or Docker, or a GitHub repo, and it answers in plain words: Safe to connect, Review before you connect, or Do not connect, with the reason. Underneath are two scores: a 0–100 trust score, backed by a signed attestation anyone can recompute offline, and an adoption score built from real usage (downloads, stars, installs). It checks for known-malicious versions, published advisories, supply-chain and dependency risk, and code-safety issues, can run the tool in a sandbox to see what it actually does, and shows an interactive trust card. Eight read-only tools cover scanning a repo, package, or live MCP server; verifying a score; checking interaction safety before you connect; looking up an identity; and getting an embeddable badge. No account, no sign-in, and nothing is modified — every call is anonymous and read-only.

Clerk

clerk.dev

Add Clerk authentication to your app with skills for setup, framework patterns, organizations, billing, webhooks, and testing. After you connect your Clerk account, inspect applications, users, organizations, domains, and instance configuration in your active workspace. With the required permissions, create applications and users, update user profiles and metadata, change instance configuration after previewing it with a dry run, and complete production setup. Private data and changes require Clerk OAuth authorization. The server does not return secret keys.

ExtensionDash: Store Rankings

ANTON KOPYLOV

Keyword rank tracking and competitor analysis for Chrome Web Store extensions and Firefox add-ons. Bring extension analytics into ChatGPT to review user growth, ratings, and search visibility over time. Research the market: search the public catalog, filter listings by users, ratings, review counts, and category, and compare competitors' descriptions. Explore gaps and use the findings to plan Chrome Web Store SEO or Firefox listing improvements. Connect an account to compare keyword rankings across dates, see which listings appear above your extension, and investigate changes in performance. Add or remove tracked extensions, keywords and competitors within your plan limits, then ask for a weekly summary. Public research works without signing in. Personal metrics and history require a connected account; results depend on the data already collected.

Domino Data Lab

Domino Data Lab

Work with the Domino Data Lab platform from ChatGPT and Codex. Skills cover workspaces, jobs, compute environments, datasets, NetApp volumes, projects, apps, model endpoints and monitoring, MLflow experiment tracking, GenAI tracing, Flows, Spark/Ray/Dask, AI Gateway, governance, tags and properties, and the Domino Python and data SDKs. Intended for data scientists and ML engineers who already have access to a Domino deployment. The skills provide guidance and code; they don't connect to Domino on their own, so running jobs or calling APIs still requires your own Domino credentials.

Build MCP Apps & Agents

Tekris Labs

Design, implement, inspect and test MCP servers and interactive MCP Apps using focused skills for tool contracts, host-aware interfaces, secure integration, debugging, visual quality and release preparation. Start with an idea, API specification or repository; work through actual coding tools available in the host. Choose real subagents when available; otherwise use a sequential workflow. Supports a useful tool-only fallback for clients without MCP Apps UI. Generated files, running tests, repository access and deployment depend on available host capabilities and user authorization; the plugin itself does not operate a hosted code-execution service.

ML Referee

Sandra Carrico

Evaluate up to 1,000 candidate features in one run. ML Referee provides independent evidence for feature selection, bounded model repair, and research prioritization without requiring a separate ablation run per feature. For machine-learning researchers and practitioners working with supervised prediction or forecasting, it checks submitted inputs, runs supported evaluations with user authorization, and retrieves results. Results are evidence within the declared evaluation setup, not guarantees of future performance or deployment readiness. An ML Referee account is required. Access requirements are described on the ML Referee website.

WeWeb

Build full-stack apps & sites

mlx-optiq Local Models

Okyasoft Pte Ltd

Guides you through running a quantized language model on your own Apple Silicon Mac with mlx-optiq. The workflow checks your Mac's memory, checks that mlx-optiq is installed (and points you to the install steps if it is not, without installing anything itself), helps you pick a model that fits, and, once you confirm the download, starts it with optiq serve as a local chat-completions endpoint. It shows how to connect a client to that endpoint, run a quick speed and quality check, and troubleshoot common problems such as running out of memory. It states the limits plainly: KV-cache quantization saves memory but can slow decoding, long contexts need memory, and quantized models should be checked on your own tasks instead of trusting speed numbers. This plugin contains instructions only; everything runs on your own machine.

Perseo Insights

Website analysis and audits