Tools

Separate skills for code and requirements review.

Prepare access to one open source with Agent Reach. It helps choose the right utility and check dependencies, for example for reading an RSS feed.

Apply for a single measurement of the local model. Check hardware support, disk, memory, and speed.

For choosing visual direction, fonts, and page composition.

A separate app for working with documents and agents. When searching materials, verify links to source fragments and answers to questions not found in documents.

README describes an editor on Next.js for App Store and Google Play, project saving, languages, and PNG export. Codex is listed among supported agents. This is declared compatibility.

A collection of DESIGN.md examples helps you choose a structure for your project’s design rules.

This skill provides Anthropic-style branding. Your own brand needs its own colours and rules.

Separate variants: library, CLI, and cloud. README describes usage with Codex, but actual connection in our research is unverified. Start with a page reading example without

Creating a team interface for a specific program.

The skill helps create a static composition such as a cover, poster or other visual asset.

Shorter answers and context handling.

Related communication processes in the ChatPlace service.

Build a Chrome extension that shows the current page title when you click its button. Manifest V3 defines its structure and permissions.

Local request gateway to selected providers.

Settings for Claude Code from publication.

Testing work through a tool from the Claude catalog.

Memory between work sessions in supported environments.

Standalone development agent, not a Codex add-on. Compare on the same task, cost, and scope of rights.

Hosting static pages and server applications.

Searching for security issues in the project's available code.

Local map of functions and connections for MCP and CLI. README describes Codex and platform choice. First build the map, then compare review of one change with and without it. Claimed

Search documentation for specific library and version.

Python tool for task and role chains. Do not count each agent as independent fact confirmation. Check input, output, repeats, and errors of each step.

AI features inside DaVinci Resolve.

Compare DeepSeek Harness and Codex on one practice task. Harness runs a separate agent environment with its own model and access settings.

Helps turn a visual reference into design rules. Behaviour and unseen details remain unknown from a single screenshot.

ECC adds a set of skills and working rules for your assistant. Consider it for a task with several stages that needs a shared approach to preparation and checking.

A set of design and interface development skills, including apple-design, mobile checks, and animation verification. README explains connection to author's experience and WWDC principles translation for we

For meaningful animation: movement should explain the action.

Build and deliver the mobile app.

A program for processing media files: trimming, assembling, converting and handling audio.

Extract page data for search and processing. Does not guarantee reading any site. Save URL, date, error, and source fragment; check incomplete results.

Voice Input in Mac Applications.

Official GSAP skill set: timelines, ScrollTrigger, React, and performance. README claims Codex support via Agent Skills. Instructions in third-party README do not replace

Skills for styling and motion for Claude Code, claude.ai, and Cowork. README warns that installation only from the repo may not include modules. Direct Codex compatibility is not

Run builds and tests when your project changes with GitHub Actions. Start with checks that already work locally.

Official collection confirms the name Bites, short practical lessons, and lists topics. All lessons, labs, rates, and regional availability are unverified. N

Select the specific Google task skill, predefine keys and read or modify permissions.

DESIGN.md stores interface rules in a text file. Google’s specification and validation tools are separate parts of the workflow.

Official documentation describes task execution and viewing changes. Cloud results must be verified before project acceptance. Jules is not a Codex plugin. GitHub connection

File Relationship Map to Quickly Find Relevant Code.

Headroom shortens material before it reaches a model. Start with a fictional long log so you can check whether a known error survives.

Standalone agent with memory and tools.

Graphics and video from HTML, CSS, and animations.

Build Three.js scene from image.

A set of approaches and commands for reviewing and improving interfaces. Choose a specific command for an observed problem.

Voice web interface with Claude Agent SDK and local bridge. README requires Claude Code with subscription, Node 20+, and Chrome/Edge. ElevenLabs is optional; a fallback uses

Anthropic's educational example for Claude Managed Agents. README clearly warns the project is unsupported, and the explanatory process consumes more tokens than specialized

Request classifier. Measure selection error and uncertainty class, then delay.

Turn your own note into a short LinkedIn post. LinkedIn Skills helps keep your voice and produces a draft for review.

Instagram analysis service: public metrics, speech, frames, and video comparison with usual author response. Service page read. MCP address https://mcp.loore.ai/mcp indicated in publi

Build an MCP server that reads practice notes from one folder. The SDK helps write the server; connecting it to your Codex is a separate check.

Extracts available file content into Markdown. That text does not prove preservation of layout, images or every table.

Clarify the task before starting implementation.

Apply for hypotheses, do not present simulation as customer survey. Model and memory keys are needed.

A set of motion principles: duration, smoothness, sequence, and feedback. README declares Codex and various animation systems. These are agent instructions, not a generato

Routing requests through a single local address.

README describes a set of research, scenario, production, and memory skills in Obsidian. MIT license. Documented manual installer for macOS/Linux and plugin commands copy.

Official developer article confirms address https://mcp.ootto.ai/mcp, speech and frame analysis, content creation, and message handling. The author of publications is also the creator

Give Codex search and reading access to official OpenAI documentation through Docs MCP. For example, find MCP configuration instructions for your client.

Combined editing process with multiple engines.

Distribution of multiple agents by tasks and budget.

Python tool for voice dialogs. Codex can help write the app; it is not a ready phone line. Separately needed are model, speech, connection, script, and delay check.

Browser control from command line when working with an agent.

Helps repeat browser actions and check results. Reading HTML does not replace an actual browser run.

Use PostHog to see how many people successfully submit a form. Start with one event so the metric is clear and testable.

Publication planning. README separates cloud and own server; own server option requires infrastructure and model key setup. First verify draft and distrib

Task templates for different models and tasks.

Edit animated scenes in a Raylight browser project through your assistant. It can read a scene, change it and show rendered frames for review.

Videos with precise typography, graphics, and repeatable scenes.

Turn descriptions on your page into a table with ScrapeGraphAI. A model helps interpret the text; check every value against its source fragment.

Extract names and prices from a small HTML file with Scrapling. Start with your own saved page to test extraction rules without requests to a website.

Creates an interface starting point from an image. The image cannot reveal unknown form submission, storage or access rules.

Search suitable skill in separate corpus.

Selecting skill text based on repeated task results.

Check third-party skill files before installation.

Workflow specifically for Codex. According to current README, solo work is normal mode; one assistant added as needed; full cycle is a separate exception. Video

Start with your own test app, save vulnerability example and retest fix.

Database, user login, and file storage.

Repeatable process from task discussion to review.

Official README: tested only on Linux; requires NVIDIA with minimum 24 GB memory, CUDA recommended 12.4. Speeds measured on H100, not a regular laptop. MIT license. Direct

Interface design rules.

Build a small desktop app: one window with a note and file saving. Tauri packages a web interface as a separate application.

Helps choose styling for a document, slides or interface. A theme does not verify content or functionality.

Turn your note into a short Threads sequence. Threads Skills helps connect the posts around one idea and remove repetition.

A set of reference skills for scenes, lighting, materials, model loading, and interaction. README header mentions Claude Code, but clone example points to another pink repository

Helps choose and describe interface decisions for a task. Check both setup and results in your own environment.

Check signs of unfinished agent work.

Checked README sections: npx skills find search, specific skill and agent installation, project/global scope. Codex is clearly among supported agents. This is a check of d

Website deployment with verifiable versions.

Set of roles for product discussion, code review, and testing.

Build one n8n workflow: turn a practice request into a clear summary. Start with manual runs and no external sending.

A helper for working with video. Check specific capabilities, setup and where files are sent before using it.

App screenshots for the store.

A local program for recognising text in images. Russian text requires the corresponding language model.

A Markdown notes editor. Links and search help organise files; creating a folder does not automatically enable AI or sync.

An audio editor for testing speech processing. Noise Reduction targets steady background noise, rather than every kind of interference.