softaworks agent-toolkit
COMMUNITYLABSCO SUMMARY
The author's own README sorts the 43 into ten categories, and the split is real: documentation and diagramming is the largest group (backend-to-frontend-handoff-docs, c4-architecture, crafting-effective-readmes, draw-io, excalidraw, frontend-to-backend-requirements, marp-slide, mermaid-diagrams, and writing-clearly-and-concisely), followed by everyday development work (database-schema-designer, dependency-updater, naming-analyzer, lesson-learned, reducing-entropy, session-handoff) and design/frontend skills (design-system-starter, mui, openapi-to-typescript, react-dev, react-useeffect). Smaller groups cover planning (game-changing-features, gepetto, requirements-clarity, ship-learn-next), workplace communication (daily-meeting-update, difficult-workplace-conversations, feedback-mastery, professional-communication), three CLI wrappers for other AI tools (codex, gemini, perplexity), and four meta-skills for building Claude Code plugins and commands themselves (agent-md-refactor, command-creator, plugin-forge, skill-judge). What's left is a genuine grab bag: one QA planner, one git-commit skill, and a utilities set running from a Datadog log/metrics CLI to a meme generator built on memegen.link.
There's no single reader this package is "for" — it's for whoever wants the specific skill they came for, not for installing the set. Thirty of the 43 run with nothing beyond the code already in front of you; ten assume some local CLI is already installed (the codex, gemini, and draw.io binaries among them); jira and perplexity go through an MCP server; and datadog-cli is the one that needs an account key on file.
READ THE FULL ANALYSIS
Read the individual skill before trusting it. Several entries make claims we can't verify from the material we have — humanizer credits its rule set to Wikipedia's own "Signs of AI writing" guide and to a separate open-source skill by a named third-party author it was forked from, and game-changing-features and gepetto both promise structured multi-step interviews whose actual output quality depends entirely on how the conversation goes, not on anything checkable in the SKILL.md text itself.
ALSO IN THIS PACKAGE
codex
Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing. Uses GPT-5.2 by default for state-of-the-art software engineering.
gemini
Use when the user asks to run Gemini CLI for code review, plan review, or big context (>200k) processing. Ideal for comprehensive analysis requiring large context windows.
perplexity
Web search and research using Perplexity AI. Use when user says 'search', 'find', 'look up', 'ask', 'research', or 'what's the latest' for generic queries.
agent-md-refactor
Refactor bloated AGENTS.md, CLAUDE.md, or similar agent instruction files to follow progressive disclosure principles. Splits monolithic files into organized, linked documentation.
command-creator
Create Claude Code slash commands. Use when users ask to 'create a command', 'make a slash command', 'add a command', or want to document a workflow as a reusable command.
plugin-forge
Create and manage Claude Code plugins with proper structure, manifests, and marketplace integration. Use when creating plugins for a marketplace, adding plugin components (commands, agents, hooks), bumping plugin versions, or working with plugin.json/marketplace.json manifests.
skill-judge
Evaluate Agent Skill design quality against official specifications and best practices. Use when reviewing, auditing, or improving SKILL.md files and skill packages.
backend-to-frontend-handoff-docs
Create API handoff documentation for frontend developers. Use when backend work is complete and needs to be documented for frontend integration.
c4-architecture
Generate architecture documentation using C4 model Mermaid diagrams. Use when asked to create architecture diagrams, document system architecture, visualize software structure.
crafting-effective-readmes
Use when writing or improving README files. Provides templates and guidance matched to your audience and project type.
draw-io
draw.io diagram creation, editing, and review. Use for .drawio XML editing, PNG conversion, layout adjustment, and AWS icon usage.
excalidraw
Use when working with *.excalidraw or *.excalidraw.json files, user mentions diagrams/flowcharts, or requests architecture visualization.
frontend-to-backend-requirements
Document frontend data needs for backend developers. Use when frontend needs to communicate API requirements to backend.
marp-slide
Create professional Marp presentation slides with 7 beautiful themes. Use when users request slide creation, presentations, or Marp documents.
mermaid-diagrams
Comprehensive guide for creating software diagrams using Mermaid syntax. Use when users need to create, visualize, or document software through diagrams.
writing-clearly-and-concisely
Use when writing prose humans will read—documentation, commit messages, error messages, explanations, reports, or UI text.
design-system-starter
Create and evolve design systems with design tokens, component architecture, accessibility guidelines, and documentation templates.
mui
Material-UI v7 component library patterns including sx prop styling, theme integration, responsive design, and MUI-specific hooks.
openapi-to-typescript
Converts OpenAPI 3.0 JSON/YAML to TypeScript interfaces and type guards. Use when generating types from OpenAPI or creating API interfaces.
react-dev
Build React components with TypeScript. Covers type-safe patterns for React 18-19 including generic components, proper event typing, and routing integration.
react-useeffect
React useEffect best practices from official docs. Use when writing/reviewing useEffect, useState for derived values, data fetching, or state synchronization.
database-schema-designer
Design robust, scalable database schemas for SQL and NoSQL databases. Provides normalization guidelines, indexing strategies, migration patterns.
dependency-updater
Smart dependency management for any language. Auto-detects project type, applies safe updates automatically, prompts for major versions.
naming-analyzer
Suggest better variable, function, and class names based on context and conventions.
reducing-entropy
Manual-only skill for minimizing total codebase size. Only activate when explicitly requested.
session-handoff
Creates comprehensive handoff documents for seamless AI agent session transfers. Use when context window approaches capacity or work session is ending.
lesson-learned
Analyze recent code changes via git history and extract software engineering lessons. Use when reflecting on work or asking 'what is the lesson here?'
game-changing-features
Find 10x product opportunities and high-leverage improvements. Use when user wants strategic product thinking or wants to find high-impact features.
gepetto
Creates detailed, sectionized implementation plans through research, stakeholder interviews, and multi-LLM review. Use when planning features that need thorough pre-implementation analysis.
requirements-clarity
Clarify ambiguous requirements through focused dialogue before implementation. Use when requirements are unclear or features are complex.
ship-learn-next
Transform learning content (like YouTube transcripts, articles, tutorials) into actionable implementation plans using the Ship-Learn-Next framework.
difficult-workplace-conversations
Structured approach to workplace conflicts, performance discussions, and challenging feedback using preparation-delivery-followup framework.
feedback-mastery
Navigate difficult conversations and deliver constructive feedback using structured frameworks. Covers the Preparation-Delivery-Follow-up model and Situation-Behavior-Impact (SBI) technique.
daily-meeting-update
Interactive daily standup/meeting update generator. Use when user says 'daily', 'standup', 'scrum update', 'status update', or wants to prepare for daily meeting.
professional-communication
Guide technical communication for software developers. Covers email structure, team messaging etiquette, meeting agendas, and adapting messages for technical vs non-technical audiences.
qa-test-planner
Generate comprehensive test plans, manual test cases, regression test suites, and bug reports for QA engineers.
commit-work
Create high-quality git commits: review/stage intended changes, split into logical commits, and write clear commit messages (including Conventional Commits).
datadog-cli
Datadog CLI for searching logs, querying metrics, tracing requests, and managing dashboards. Use when debugging production issues or working with Datadog observability.
domain-name-brainstormer
Generates creative domain name ideas for your project and checks availability across multiple TLDs (.com, .io, .dev, .ai, etc.).
humanizer
Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written.
meme-factory
Generate memes using the memegen.link API. Use when users request memes, want to add humor to content, or need visual aids for social media.
web-to-markdown
Converts webpage URLs to clean Markdown by calling the local web2md CLI (Puppeteer + Readability), suitable for JS-rendered pages.
jira
Use when the user mentions Jira issues (e.g., "PROJ-123"), asks about tickets, wants to create/view/update issues, check sprint status, or manage their Jira workflow. Triggers on keywords like "jira", "issue", "ticket", "sprint", "backlog", or issue key patterns.
agent-ascii-ui-mockup-generator
Visualize UI concepts through ASCII mockups before implementation. Creates multiple ASCII mockup variations for dashboard layouts, forms, and other UI components.
agent-codebase-pattern-finder
Find similar implementations, usage examples, or existing patterns in the codebase. Provides concrete code examples based on what you're looking for.
agent-communication-excellence-coach
Review communication drafts or prepare difficult conversations. Provides email refinement, tone calibration, roleplay practice, and presentation feedback.
agent-general-purpose
Default agent for handling complex, multi-step tasks with automatic delegation capabilities.
agent-mermaid-diagram-specialist
Mermaid diagram specialist for creating flowcharts, sequence diagrams, ERDs, and architecture visualizations.
agent-ui-ux-designer
Expert UI/UX design critic and advisor who provides research-backed, opinionated feedback on interfaces. Pushes back on bad ideas and cites sources.
command-codex-plan
Create a detailed implementation plan using Codex 5.2 with high reasoning.
command-compose-email
Draft a professional email using the What-Why-How framework. Use when composing emails to colleagues, stakeholders, or leadership.
command-explain-changes-mental-model
Build a mental model of changes by splitting them into smaller logical chunks.
command-explain-pr-changes
Generate comprehensive markdown summary of pull request changes with visual impact analysis.
command-sync-branch
Sync feature branch with the latest main branch.
command-sync-skills-readme
Sync root README.md with current skills inventory from skills/ directory.
command-viral-tweet
Transform a raw tweet idea into an optimized viral post for X. Analyzes engagement drivers and generates 3 optimized versions with posting strategy.
ascii-ui-mockup-generator
Use this agent when you need to visualize UI concepts through ASCII mockups before implementation.
codebase-pattern-finder
codebase-pattern-finder is a useful subagent_type for finding similar implementations, usage examples, or existing patterns that can be modeled after. It will give you concrete code examples based on what you're looking for!
communication-excellence-coach
PROACTIVELY use when reviewing communication drafts or preparing difficult conversations. Provides email refinement, tone calibration, roleplay practice, and presentation feedback with actionable suggestions.
general-purpose
Default agent for handling complex, multi-step tasks with automatic delegation capabilities
mermaid-diagram-specialist
Mermaid diagram specialist for creating flowcharts, sequence diagrams, ERDs,
ui-ux-designer
Expert UI/UX design critic and advisor who provides research-backed, opinionated feedback on interfaces. Use when you need honest assessment of design decisions, want to avoid generic "AI slop" aesthetics, need evidence-based UX guidance, or want distinctive design direction grounded in actual user behavior research.
WHAT'S INSIDE
43 showing · 43 totalNothing else to set up — install it and go.
agent-md-refactor
The instructions file an AI coding assistant reads has grown into a wall of text; this splits it into a short main page that links out to one note per topic.
backend-to-frontend-handoff-docs
A feature's server-side work is finished and the people building the screens now need to know how to use it; this writes that document for them, saved in the repo and detailed enough that nobody has to interrupt you.
c4-architecture
Architecture diagrams usually live in a slide deck nobody updates; these are drawn from the codebase itself and saved as text files next to it, so they can be changed like code.
codex
Hands a job over to OpenAI's Codex tool — reviewing a codebase, reworking it, or making the edits itself — with the freedom it is given matched to how risky the job is.
command-creator
Turns something you keep asking your coding assistant to do into a named shortcut you can run in one word, kept either with the project or across everything you work on.
commit-work
Committing a day's work in one lump leaves a change nobody can review; this goes through the edits, decides where the boundaries are, and puts each group in on its own with a message that explains it.
crafting-effective-readmes
The front page of a code project — what it is, how to get it running — written for whoever actually opens it: a stranger on the internet, a new teammate, or yourself in six months.
daily-meeting-update
Nobody remembers what they did yesterday: this interviews you, offers to go through yesterday's actual activity first to jog your memory, and writes up the update the morning meeting expects.
database-schema-designer
Say what the software has to keep track of — customers, orders, whatever it is — and get the database design back: the tables, how they connect, and what keeps them quick to search.
datadog-cli
Datadog keeps a record of everything running software does; this searches that record from your terminal — the errors, the numbers, the trail one failing request left — while you are working out what broke.
dependency-updater
The outside libraries a project leans on fall behind and start carrying known security holes; this works out what the project is built in, moves it forward as far as is safe, and untangles the installs that break.
design-system-starter
Sets the shared rules a product's interface is built from — its colours, its type, its spacing, how the pieces stack up and what it takes to stay usable for everyone — so every screen looks like the same product.
difficult-workplace-conversations
Walks you through one hard talk at work from both ends: working out what you actually want and whether you are calm enough to say it, then what to do in the days afterwards so the problem does not come back.
domain-name-brainstormer
Naming a new project usually means hours of guessing and typing names into a registrar; this comes up with the candidates and checks which are still free in the same pass.
draw-io
Builds and tidies draw.io diagrams — arrows that land where they should, even spacing, the right cloud icons — and exports the finished thing as an image.
excalidraw
Excalidraw drawings are enormous files of raw data, so a separate helper opens them instead of your main session: you say what should change, and the diagram comes back edited.
feedback-mastery
How to handle the conversation at work you have been putting off — telling someone their work is slipping, or sorting out a clash between two people.
frontend-to-backend-requirements
Heads off the back-and-forth where a screen gets built against data that does not exist yet, by writing down what it needs first, in plain terms.
game-changing-features
A brainstorming session with one rule: not what to fix next, but what would make this product ten times more worth using.
gemini
Hands your code or your written plan to Google's Gemini for a second opinion, which helps most when there is more material than one sitting can cover.
gepetto
Does the homework that usually gets skipped before a software feature is built — what already exists, what you actually want, and where the plan falls down.
humanizer
Rewrites text that reads like a chatbot wrote it — flat, over-polished, every sentence the same length — into something with an opinion, a rhythm and a person behind it.
jira
Jira tickets managed in plain language — look one up, raise a new one, move it to the next stage — over whichever Jira connection your machine already has.
lesson-learned
Looks at the code you just wrote and names the one thing it taught you — a mirror held up to your own work, not a lecture on best practices.
marp-slide
Builds a slide deck out of plain text in one of seven ready-made visual styles — you write the words, the design is already decided.
meme-factory
Makes a meme from one of a hundred-plus familiar picture templates and your own two lines of text, and hands back a link to the finished image.
mermaid-diagrams
Draw a diagram by typing it out rather than dragging shapes around, so the picture lives alongside the code and gets edited the same way.
mui
Building a React interface with Material UI version 7 — which components to use, how its built-in styling works, and what version 7 changed from version 6.
naming-analyzer
Programmers name everything in their code, and a bad name quietly misleads whoever reads it next; this goes through those names and says which ones are wrong and what to call them instead.
openapi-to-typescript
Reads the file that describes a web service and writes out the matching TypeScript code, so nobody hand-copies every field name into their own types.
perplexity
Searches the live web through Perplexity — either a short list of links worth reading, or a written answer built from what is online right now.
plugin-forge
Sets up a new add-on for Claude Code so the folders, the settings files and the listing all line up the way the tool expects.
professional-communication
Workplace writing that actually lands — an email, a chat message, a status update, a meeting agenda — each one shaped for the person who has to read it and act on it.
qa-test-planner
Writes the paperwork of software testing — what gets tested, in what order, and what to file when something breaks.
react-dev
Writing React screens in TypeScript so the mistakes show up while you type rather than after the page is live.
react-useeffect
React's own advice about the useEffect hook, which is mostly about the times you should not reach for it and what to write instead.
reducing-entropy
One rule applied to every change: is there less code in the end than there was before? Effort spent does not count, only what is left behind.
requirements-clarity
Asks the questions a vague feature request leaves unanswered — what exactly, for whom, and what happens when it goes wrong — until there is enough to write it down properly.
session-handoff
A written handover for whoever picks the work up next — where things stand, what was decided and why, and what to do first.
ship-learn-next
Turns a video or article you have just taken in into the first small thing you will actually make this week, and the few that follow it.
skill-judge
Marks a skill's instruction file the way an examiner marks a paper — eight scores, a letter grade from A to F, and the three changes worth making first.
web-to-markdown
Saves a web page as a plain text file with just the article in it, no menus or ads, including pages that only appear once a real browser has run them.
writing-clearly-and-concisely
The plain rules of good writing — say it straight, in the fewest words that still carry it — applied to anything you have written for a person to read, from an error message to a report.
HOW TO GET IT
npx skills add softaworks/agent-toolkitnpx skills add softaworks/agent-toolkit --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.
/plugin marketplace add softaworks/agent-toolkit/plugin install codex@agent-toolkit/plugin install gemini@agent-toolkit/plugin install perplexity@agent-toolkitTyped inside the agent's own prompt, not in a terminal. The marketplace lists 56 plugins; three are shown. Every one installs the same way, with its own name before @agent-toolkit.