Language: English | Tiếng Việt | 中文
A collection of method-driven skills for people who use AI to make decisions, run projects, analyze evidence, create documents, and produce finished work. These skills are designed for active collaborators: you provide the real context and remain responsible for the decision; the agent supplies structure, analysis, and quality checks.
Install the full collection:
npx skills add tronghieu/agent-skillsOr install as a Claude Code plugin:
/plugin marketplace add tronghieu/agent-skills
/plugin install agent-skills@tronghieu-skills
Then invoke a skill in your agent:
/market-researcher Assess the market for this product idea in Vietnam.
Most agent tools use slash commands. Codex and ChatGPT use $; in those tools, replace the leading / with $.
| Your job | Start with |
|---|---|
| Generate or challenge ideas | brainstorm-coach, critical-thinking |
| Diagnose a recurring problem | problem-solver |
| Understand customers or markets | design-thinking, market-researcher |
| Decide what to build and deliver it | product-manager, project-manager |
| Run software sprints and improve team process | scrum-master |
| Analyze data or long documents | data-scientist, deep-reader |
| Make a high-stakes business decision | strategy-board |
| Create documentation, prompts, slides, or fiction | diataxis-writer, system-prompt-creator, slidewright, fiction-studio |
| Learn through guided questions | socratic-questor |
| Compare CVs with a job description | cv-scorer |
| Understand what an autonomous coding run did, or why it stalled | bmad-run-inspector |
| Test a running app by hand and prove what you saw | manual-tester |
For anyone running bmad-loop, an autonomous coding orchestrator, who needs an honest read on a run — watching one in progress or investigating one that finished, failed, or paused overnight. Only useful inside a project that already runs bmad-loop.
npx skills add tronghieu/agent-skills --skill bmad-run-inspectorRead the bmad-run-inspector guide
For founders, product teams, marketers, creators, and anyone who needs to generate options before evaluating them.
npx skills add tronghieu/agent-skills --skill brainstorm-coachRead the brainstorm-coach guide
For decision-makers, analysts, and writers who need to inspect the claims, evidence, assumptions, and logical gaps in a document.
npx skills add tronghieu/agent-skills --skill critical-thinkingRead the critical-thinking guide
For recruiters and hiring managers who want a consistent, rubric-based comparison of CVs against a job description. It supports human review; it does not make hiring decisions.
npx skills add tronghieu/agent-skills --skill cv-scorerFor data practitioners and decision-makers who need defensible exploration, statistical analysis, predictive baselines, or a review of an existing analysis.
npx skills add tronghieu/agent-skills --skill data-scientistFor researchers, students, analysts, and serious readers working with books, theses, papers, or other documents of roughly 50 pages or more.
npx skills add tronghieu/agent-skills --skill deep-readerFor product, design, and innovation teams that can collect real user evidence and want a disciplined path from research to tested concepts.
npx skills add tronghieu/agent-skills --skill design-thinkingRead the design-thinking guide
For technical writers, documentation owners, developer advocates, and teams improving tutorials, how-to guides, references, explanations, or knowledge bases.
npx skills add tronghieu/agent-skills --skill diataxis-writerRead the diataxis-writer guide
For authors developing prose fiction—from a first premise through outline, drafting, revision, and continuity review.
npx skills add tronghieu/agent-skills --skill fiction-studioFor anyone who has to test a running app by hand and wants findings that survive scrutiny: exploratory bug hunts, verifying a fix, smoke passes, UX or localization review with ranked improvements, probing an AI assistant's answers, and writing manual test plans that skip what automated tests already prove. Every finding is checked against the database, not just the screen. Needs a local dev stack.
npx skills add tronghieu/agent-skills --skill manual-testerFor founders, product teams, strategists, consultants, and analysts who need cited desk research on market size, competitors, demand signals, or trends.
npx skills add tronghieu/agent-skills --skill market-researcherRead the market-researcher guide
For operators, technical leads, and business owners who need to verify a root cause before choosing a fix.
npx skills add tronghieu/agent-skills --skill problem-solverFor working product managers, product leads, and founders deciding what to build, how to measure it, and how to launch it.
npx skills add tronghieu/agent-skills --skill product-managerRead the product-manager guide
For project managers, PMOs, delivery leads, and team leads planning, tracking, de-risking, or recovering a real project.
npx skills add tronghieu/agent-skills --skill project-managerRead the project-manager guide
For Scrum Masters, engineering managers, tech leads, product owners, and software teams that need focused sprint planning, honest sprint health, visible impediments, and retrospective follow-through.
npx skills add tronghieu/agent-skills --skill scrum-masterFor speakers, educators, consultants, and founders building web-based slides for a live presentation.
npx skills add tronghieu/agent-skills --skill slidewrightFor learners and teachers who want to build or test understanding through a guided Socratic dialogue.
npx skills add tronghieu/agent-skills --skill socratic-questorRead the socratic-questor guide
For founders, executives, and strategy leaders working through consequential choices such as market entry, investment, build-versus-buy, pricing, or transformation.
npx skills add tronghieu/agent-skills --skill strategy-boardFor AI product builders, developers, automation teams, and power users creating testable system prompts and custom instructions.
npx skills add tronghieu/agent-skills --skill system-prompt-creatorRead the system-prompt-creator guide
Download a skill directory or its packaged .zip, then copy the extracted directory to:
- Project scope:
.agents/skills/or.claude/skills/ - User scope on macOS/Linux:
~/.agents/skills/or~/.claude/skills/ - User scope on Windows:
%USERPROFILE%\.agents\skills\or%USERPROFILE%\.claude\skills\
The exact location depends on the agent you use.
Contributions are welcome. To propose a change:
- Fork the repository and create a focused branch.
- Place each skill in
skills/<skill-name>/with aSKILL.md, any required scripts or references, and user-facing README files. - Test the skill with representative prompts, including at least one difficult or failure case.
- Rebuild
skills/<skill-name>.zip. - Open a pull request explaining who the skill is for, when it should activate, what you tested, and any known limitations.
For substantial new skills, opening an issue first can help confirm scope and avoid overlapping work. See AGENTS.md for skill-authoring conventions.
This repository follows the open Agent Skills specification. Compatibility depends on each host tool's implementation of that specification. The skills are tested on Claude and DeepSeek models.
MIT