Browse AI employees

Find AI for your task. Compare prices, required tools and deliverables before choosing.

11 AI employees

OKZAQ

Dify Design Templates Collection

This is a collection of templates for Dify's Chatflow to be "decided by the design document before it is set up on the screen". We have put together a set of design skills for Claude Code and three completed design documents (with fictional subject matter and test data) made with that skill. ■ For such troubles· When the number of nodes exceeds 20, the line connection and the reading and writing of variables can not be tracked. · Start assembling without deciding on specifications, and rebuild later After passing the DSL (YAML), the other party's version of Dify could not be loaded ■ Contents· Design Skill (SKILL.md): If you tell Claude Code that you want to build a chatbot with Dify, you can selectively listen to the requirements and create a design document (Markdown) and test data · Design guidelines by node: 12 different types of nodes, 4 design patterns, 6 anti-patterns· Improvement pattern collection: 6 patterns, such as rewriting the question before the search and reordering the search results 3 completed design documents - Internal IT help desk (25 nodes).If not, email the person in charge to take over) - Restaurant reservation · FAQ (6 nodes.Minimum configuration for the first piece to be assembled)- Job Matching (36 nodes.How many round trips do you need to listen to the conditions and propose) The design document even includes checklists for writing and reading conversation variables, flow diagrams, full text of node-by-node settings and prompts, test perspectives, and procedures for assembling on-screen.On the Dify screen, assemble the design documents in order from the top. ■ What's not included · DSL file that moves as soon as it is loaded (it is assembled on the screen by viewing the design document) · Dify's environment and AI model usage fee (the buyer's environment/contract will be used) · Content for Workflow-style apps that do not have conversations■ What you need · Dify (The design document is assumed to be the cloud version.You can also use the design concept in the self-hosted version) Claude Code corresponding to skills when using design skills.3 completed design documents can be used without skill ■ Support· For 30 days after delivery, we will respond to questions about usage and consultation of problems on the transaction page

Hire this AI employee

¥4,800one-time

note × SNS automatic operation agent (Claude Code skill set)

Social mediaClaude Code Skill

atommy

note × SNS automatic operation agent (Claude Code skill set)

it is a set of agents that can be moved with one command of Claude Code, from generating draft articles to creating X and Threads posts and Buffer reservations. [Main commands] · Generate a draft of the note article from the clause in/draft-sources/(header image is also automatically generated)·/sns-post/threads-post - Generate two draft posts for X and Threads from the draft · Buffer reservation by comparing/choose-post-A · B2 proposal ·/approve — Action pending approval (note publishing/social media posting) is processed together · Read/scheduler-SCHEDULE.md to automate today's tasks [Comparison selection of A/B posts]For each post, we propose two ideas: number-oriented and question-oriented.The pattern you select will be recorded and used in conjunction with your weekly metrics analysis to improve your next post. [Automation of material collection]if you set the theme and keywords for research/RESEARCH.md, the scheduler will collect and save the material to sources/every day.If you have a lot of material, a draft will be automatically generated every week, and the flow to post it in the note will automatically turn. Estimated time requiredThe confirmation process (draft confirmation, note public approval, social media post approval) per article takes about 5 minutes.Approximately 1 hour per month at a pace of 3 per week (5 minutes x 3 per week x 4 weeks).Everything else - gathering material, generating drafts, generating header images, creating posts - is done automatically. [Safety design]for irreversible actions such as note publishing and social media posts, the agent does not directly perform them, but loads them into the approval queue (queue/pending.md). You can check them together in/approve. [Do not] · Publish notes without approval · Do not post on social media· Do not fabricate testimonials, achievements, or numbers (use only the information written in sources/) [Necessary environment] Claude Code (paid plan or API key) · Buffer account (free plan possible) note account + note-scripts (Playwright automatic posting scripts) · Poe API key (recommended) or Anthropic API keyPoe Code (mit License) [Deliverables] 11 Claude Code skills (.claude/commands/) /draft/edit/sns-post/threads-post/choose-post/approve/spawn-manager/scheduler/draft-learn/sns-learn/research-update SCHEDULE.md (Natural Language Scheduling File)research/RESEARCH.md (spoiler collection theme keyword configuration file) Manager prompt (for multi-agent configuration) · Scripts (Scheduler · Buffer integration · Material collection · Dashboard generation · note posting · image generation) · Dashboard template (data/dashboard.json)Setup instructions (README.md) Support We will respond to your questions and problems on the transaction page for 30 days after delivery.The operation may change due to changes in specifications of Buffer, each SNS, and Claude Code.

Hire this AI employee

¥14,800one-time

MarketingClaude Code Skill

Daniel

SEO Growth / SEO diagnosis skill set

The SEO Growth Skills Collection is a set of 28 SEO skills for AI agents that can load SKILL.md such as Claude Code. Ask agents to “Audit my site”, “Why are my clicks down?”, “Why aren't I indexed?”, “Was this change effective?", we will follow the established procedure and summarize the results. [What you can do] · Site-wide audit (first check the stage of the site and perform the necessary checks in order) · Separation of traffic drops (measurement problems, seasonality, updates, impact of AI functions, actual losses)· Check index, robots.txt, sitemap, canonical, redirect · Comparison before and after the measure (Compared to the page that has not changed, we will tell you if we can not judge it) Procedures for determining updates, integrations, and deletions (deletions are not recommended out of the blue) · Checking for signs that the template page is likely to be exposed to the spam policyDetection of old years and discrepancies between pages · Searches for new content candidates, quality, internal links, legal and regulatory checks [How far can you go] Each skill will clearly indicate what you have checked, what you have not checked, and how confident you are.The results are for human judgment and we do not edit the site.Statements about Google's specifications will be marked with a confirmation date, and those that have not been officially confirmed will be marked as "unconfirmed".The reference value that Google does not publish is clearly stated as our initial value. [What the buyer needs] AI agents that can read SKILL.md (such as Claude Code)· If there is an export or API of Search Console, it will improve the accuracy of motion analysis, index checking and effect measurement (even if it is not, each skill will show an alternative procedure) · No paid SEO tools or accounts required.The 8 skills come with a small script that runs in Node.js (optional).[What you can't do] · Rank, access, and sales are not guaranteed. · Not legal, tax, financial, or medical advice (legal and regulatory skills are a risk awareness checklist). · It is not related to Google, Bing, or Employee Store, and is not officially approved.Accuracy verification is not performed on actual site data. [Language] The body of the skill is in English.If you want your answers to be in Japanese, ask your agent to answer in Japanese (we do not verify the output quality in Japanese individually). [Update]We plan to periodically review the specifications of Google and AI models as they change, and provide an updated version for buyers. [License] One buyer (or one business) is available on your site and that of your customers.It cannot be redistributed or resold to a third party.

Hire this AI employee

¥4,560one-time

10 Geo skills to be cited and recommended by AI (with mit/Japanese guide)

MarketingClaude Code Skill

Howseen AI

10 Geo skills to be cited and recommended by AI (with mit/Japanese guide)

10 agent skills to ensure your brand is cited and recommended by AIs like ChatGPT, Perplexity, and Gemini.It's not about Google's ranking, it's about getting an AI's answer. by Howseen AI You don't need an API key or code.When you ask an agent to specify what you want to do with your site, the skills that match are automatically loaded.Claude Code, Cursor, Codex, and other agents that support the Agent Skills format. This skill set is open source (mit license) and the code is published for free.This is not an exclusive code to buy, but a ready-to-use package, a guide to getting started in Japanese, and post-purchase support. Contains Skills Content Page · llms-txt: Create llms.txt for AI to understand the site.Add what you can and can't do· content-capsule: rewrite the page into a short sentence with "Answer first" that is easy for AI to quote as it is · schema-builder: Create structured data (JSON-LD) that the operator tells AI who Entity-coverage: Compare the topics and pages that experts would touch on, and list the missing points· faq-schema: Make the page into a question and answer pair, and also create the JSON-LD of the FAQPage How to see from technology and AI · crawler-check: Check whether AI crawlers such as GPTBot, ClaudeBot, PerplexityBot can read the site from robots.txt bing-check: Check if the ChatGPT search is on Bing that you rely on· Retrievability: Score on a scale of 0 to 100 whether the AI is easy to incorporate the content of the page into the answer, and list the points to correct · Buyer-questions: Collect the questions the buyer is asking AI from public discussions citation-check: Find out if ChatGPT, Perplexity, and Gemini are already citing their company, and who is citing them insteadExample of how to use it Specify the domain and ask for it, such as “Check if your site can read the AI crawler.” Note The result of each skill is the situation as of the day it was performed.AI answers vary from week to week and from service to service. The seller has not yet confirmed how the site works in Japanese.Introduction process (for details, please refer to the accompanying guide in Japanese) 1. Copy the skills folder to your project's .claude/skills/or to ~/.claude/skills/for all projects 2. Designate your site for the agent and ask them to do what they want Deliverables Set of 10 skills (zip)-Japanese introduction guide Full mit License If you have any post-purchase questions, please send them to the seller from the trading room.

Hire this AI employee

¥12,500one-time

Skill in creating motion design videos with Claude Code (with mit/Japanese guide)

MarketingClaude Code Skill

Howseen AI

Skill in creating motion design videos with Claude Code (with mit/Japanese guide)

The skill that lets Claude Code create motion design videos with just the code. by Howseen AI Don't use After Effects, Remotion licenses, or video editing apps.Make a move in HTML, export one frame at a time in Playwright, and make a video in ffmpeg.Change the scene to match the beat of the music, and place the sound effect at the peak of the sound. This skill is open source (mit license) and the code is published for free.This is not an exclusive code to buy, but a ready-to-use package, a guide to getting started in Japanese, and post-purchase support. What you can do Ask for a video in Claude Code or type/motion-design to proceed from configuration to export· How to proceed: Check the material and requirements, assign a scene to the music shot, check with 4 still images, export the production, detect flicker, insert sound, and complete · Motion blur that takes multiple shots of a frame and overlaps it, and write it out at 60fps · Align music and sound effects, and set the volume to -14LUFS· Music, sound effects, and photos can be searched and obtained from free material sites (Mixkit, Pexels, Unsplash, etc.) · "Remake" mode to remake existing launch videos one frame at a time with your own brand When displaying sample numbers on the screen, include rules to avoid displaying them differently from the facts, such as specifying "Example data"Suitable for · I want to create a short video for product introduction and launch in-house without outsourcing Developers and designers who already use Claude Code Required environment Claude Code · Python (playwright, imageio-ffmpeg, numpy, pillow) and Chromium for Playwright Introduction process (for details, please refer to the accompanying guide in Japanese)1. Copy the Skills folder to the Claude Code Skill Locations 2. Insert Python libraries and Chromium 3. Ask for a video in Claude Code or type/motion-design Deliverables A set of skills (SKILL.md and auxiliary script) and a sample video project (zip) -Japanese introduction guideFull mit License If you have any post-purchase questions, please send them to the seller from the trading room.

Hire this AI employee

¥12,500one-time

Back officeClaude Code Skill

calez

Fumi / Email Log

This AI converts saved email threads (.eml / .mbox / pasted text) into an action log. It does not generate summaries or draft replies. Instead, it lists confirmed items, who took on what tasks and by when, pending requests that have not been accepted, conflicting deadlines for the same deliverable, and unanswered questions—all with direct quotes from the original text and message numbers. As a process specific to email, it automatically removes history quoted in replies. Even if the same commitment is quoted in four replies, it is counted as a single commitment. The number of lines removed is also clearly noted in the log. Forwarded requests are retained as part of the body text rather than as history. This is because deleting the forward would cause the request itself to disappear. The output consists of six fixed sections: Decisions, Agreements, Pending Requests, Inconsistencies, Unanswered Questions, and Extraction Notes. Each item includes a quote from the original text and a message number; the quotes are cross-referenced with the original text using the included verification script. Items with non-existent quotes are deleted. The only requirement is Claude Code (or Codex). Python 3.9 or higher is optional; if it is not available, “UNVERIFIED QUOTES” will be displayed in the log. There are no external APIs, account registrations, or communication during execution. It does not connect to IMAP, Gmail, or Outlook; it reads only saved files. There is no email sending functionality. What it cannot do: compose replies, summarize content, read attachments, merge multiple threads, analyze sent emails, or determine the assigned person. CC recipients are not treated as the assigned person. Requests remain as “pending requests” until a reply confirming acceptance is received. Supported languages are Japanese and English. For Japanese threads, the log is created in Japanese, and quotes are not translated. Since analysis and quote matching are performed on the Python side, the log format is preserved even if the model changes. Since extraction accuracy depends on the model, you can verify operation during setup using the included samples and the `tests/golden.md` file.

Hire this AI employee

¥7,800one-time

Back officeClaude Code Skill

calez

Sumi / Quote Review

This AI checks the internal calculations in a single quote or invoice. It supports text PDFs, scanned PDFs, CSV files, and pasted tables, and verifies—complete with calculation formulas—whether the “quantity × unit price,” subtotal, discount, sales tax, and total for each row match. It also handles mixed tax rates of 10% and 8%, as well as rounding of fractional amounts on a line-by-line basis. There are four output types: PASS, PASS WITH WARNINGS, FAIL (Do not send), and HOLD (Please verify the numbers), each accompanied by the underlying calculation process. The system does not create, modify, or send documents. Scanned PDFs are read twice using two different methods, and the system automatically verifies that all items match down to the single-digit level. Even if the values match, the system will not issue a final result until you verify the numbers against what is displayed on the screen; it will stop at “HOLD.” This is because reading errors and calculation errors may appear identical in the output. The only requirement is a Claude Code (or Codex). Python 3.9 or higher is optional; if it is not available, the report will display “DEGRADED” (unverified). There are no external APIs, account registrations, or communication during execution. No API keys are exchanged. What it cannot do: accounting or tax-related judgments; evaluating the reasonableness of prices; reading handwritten characters or stamped numbers; scanning images with a resolution lower than 180 dpi; documents containing a mix of two currencies; or verifying discounts that span multiple tax rates. The report will explicitly state that any of these are outside the scope of the service. Supported languages are Japanese and English. Japanese documents are inspected as-is; quoted text is not translated. Since calculations are performed on the Python side, the basis for the assessment remains unchanged even if the specifications of the model or tools are modified. Only the data-extraction process may be affected; in such cases, please re-execute the scan path verification procedure. Samples and expected output (tests/golden.md) are included, allowing you to verify functionality yourself during setup.

Hire this AI employee

¥5,800one-time

Social mediaClaude Code Skill

OKZAQ

Weekly Social Media Management Agent

This is a complete set of agents for Claude Code that simplifies your weekly social media posts to “just reading and approving proposals.” We’ve made the system that the seller uses every week for their X / Threads account available for anyone to use. ■ What It Does ・Initial Setup: Based on a conversational consultation and an analysis of your past posts, it creates a “playbook” that outlines your voice, themes, posting rules, and posting times. ・Weekly: Drafts a week’s worth of posts based on the playbook → Automatically checks character count, dates/times, and duplicates → proposes them via chat → only approved posts are scheduled in Buffer ・Tailors posts for X and Threads (sets length, hashtags, and posting times by channel) ■ What It Doesn’t Do ・It does not schedule or post without your approval ・It does not fabricate personal stories, achievements, or statistics. It uses only the notes you’ve written as primary sources (This design prevents the account’s credibility from being undermined by AI-generated personal stories.) ■ For People Like This ・Sole proprietors, small business owners, and content creators who use or want to try Claude Code ・Those who want to keep up with social media but find creating weekly posts burdensome ・Those concerned that letting AI write results in “AI-sounding” text ■ Requirements ・Claude Code (paid plan or access via API key) ・Python 3.9 or higher (no additional installation required) ・Buffer account (free plan is acceptable; the free plan allows up to 10 scheduled posts per channel) ■ Deliverables ・Two Claude Code skills (one for initial setup, one for weekly operations) ・Playbook template and sample entries ・Buffer scheduling tool (Python) and setup guide ■ Important Notes ・Does not support posts with images (text-only posts) ・The person approving the post is responsible for the final review and publication of the post content ・Functionality may change due to updates to Buffer or the specifications of each social media platform ■ Support ・For 30 days after delivery, we will respond to questions about usage and reports of issues via the transaction page

Hire this AI employee

¥9,800one-time

/remove-ai-slop: Claude Code skill for figuring out the AI-like nature of screens and words

Engineering & dataClaude Code Skill

slashskills

/remove-ai-slop: Claude Code skill for figuring out the AI-like nature of screens and words

It's a Claude Code skill that identifies patterns that often look like they were created by AI, reports them in order of priority, and makes concrete suggestions before and after fixing them from the site or app screen and text. by slashskills It's not just the code you see.Check the actual screen and only point out the evidence.For example, a sales pitch like an unfounded “AI-driven” statement, a phrase that remains tentative, a button with overlapping meanings, or a structure in which every page looks the same type. This skill is open source (mit license) and the code is published for free.This is not an exclusive code to buy, but a ready-to-use package, a guide to getting started in Japanese, and post-purchase support. What you can do · Examines the screen and words, ranks the points by magnitude of impact, frequency, certainty, and effort to correct, and reports· Pointing out is judged in three stages: things that should be clearly erased, things that should be fixed if there is no reason, and things that depend on the context For each indication, provide the applicable location, the words before the correction, the proposal after the correction, and the basis · Any unsubstantiated claim should be cited, narrowed down, specified as unconfirmed, or truncatedIf you do not know the intention of the product, do not make words without permission and return "judgment is necessary" Designs that should be left intentionally are also listed as "leftovers" - Fix it only if asked to do so, and then display the screen again after fixing it to confirm it Suitable forSites and screens created with AI look the same and are troubled Before publishing, I want to inspect the screen and text with the eyes of a third party Required environment Claude Code (Operation confirmed.It is a format that works with Codex and Cursor, but it has not been verified)The state in which you can use the files, terminal commands, and browser of the project you want to investigate Introduction process (for details, please refer to the accompanying guide in Japanese) 1. Copy the Skills folder to the Claude Code Skill Locations 2. Type/remove-ai-slop in Claude Code or write down what you want to do DeliverablesSkill set (SKILL.md, zip) -Japanese introduction guide Full mit License If you have any post-purchase questions, please send them to the seller from the trading room.

Hire this AI employee

¥3,200one-time

Engineering & dataClaude Code Skill

Shun

Claude Codex Development Operations Kit

This is a macOS operation kit for developers who use Claude Code and Codex.It aggregates the number of session events, checks the status of the Git repository, executes time-limited commands, checks confidential information patterns before publishing, and creates daily Markdown reports.[Deliverables] Zip summarizing Python source, configuration example, setup script, introduction guide, operation Skill, and test.Buy it out of stock and set it up on the buyer's own Mac. [Required environment] macOS, Python 3.9 or later, Bash, Git.The kit itself does not call a paid API.Please prepare your own agreement for the use of Claude and Codex. [Scope] The aggregation of the number of uses is not the aggregation of the billing amount.Confidential information inspection is based on pattern matching and does not guarantee leak prevention.Scheduled execution is disabled by default.Initial setting agency and monthly operation are not included.The code is the mit License.

Hire this AI employee

¥9,800one-time

Skill set to crush the silent failure of automation

Engineering & dataClaude Code SkillAI-assisted service

Shun

Skill set to crush the silent failure of automation

The process being turned unattended ends with empty contents with an exit code of 0. Claude Code skill set for detecting this elusive failure. Pack of 12. It is written only based on accidents that you actually stepped on.Such as the trap of how to write a log exit code that always reaches 0, the fake success of browser automation, and the detection of a state in which it is not possible to post with the intention of posting. Place it in the Claude Code Skill Depot and use it.

Hire this AI employee

¥9,800one-time

Browse AI employees

Browse AI employees

Ask AI if it fits your work.