IELTS Buddy: Local Agent Skills for Language Test Prep

Where this fits

You're a developer prepping for IELTS while juggling agent builds. Instead of toggling between generic chat prompts and web apps, you want your local AI agent (Codex, Claude Code, or Cursor) to act as a dedicated tutor. IELTS Buddy delivers that as a set of installable agent skills, each tuned for a specific task—from essay review to mock tests—so your agent can guide your study consistently without you re-writing prompts every session.

Why agents benefit

• Task-specific skills: Each skill (e.g., ielts-writing-review, ielts-speaking-coach) carries its own instructions, rubrics, and output formats, so the agent doesn't guess how to evaluate your work.
• Composable workflow: Install only what you need or the full set; the agent picks the right skill based on your request, like "review this essay" or "run a speaking mock."
• Local and private: All processing happens in your existing agent environment—no sending your practice responses to a third-party service.
• Easy updates: Skills are versioned via GitHub and can be refreshed with a single CLI command, keeping feedback aligned with the latest IELTS criteria.
• Multi-agent support: Works across Codex, Claude Code, Cursor, and WorkBuddy, so you can switch tools without losing your study setup.

Practical scenarios

Essay review with rubric-based feedback

You paste a Task 2 essay and say "use $ielts-writing-review to grade this." The agent applies the official band descriptors, highlights coherence, lexical resource, and grammar issues, and suggests targeted revisions.

Speaking mock with instant assessment

You record or type your answer to a cue card. The speaking coach skill simulates a real examiner, times your response, and gives feedback on fluency, pronunciation, and topic development—all within your agent chat.

Daily vocabulary and reading/listening drills

Instead of a generic word list, the vocabulary coach skill generates context-based exercises from your weak areas. You can also run a timed reading passage or listening transcript review, with the agent explaining correct answers and distractors.

Add it to your agent workflow

Install via the skills CLI (requires Node.js 18+). First, list available skills:

npx skills@latest add Jobo16/ielts-buddy --list

Install the full set globally:

npx skills@latest add Jobo16/ielts-buddy --skill '*' --global --yes

Then just describe a task: "Use $ielts-speaking-coach to practice Part 1 questions about work." The agent will invoke the skill and follow its built-in coaching flow.

Compare before adopting

Before committing, evaluate the quality of the feedback each skill produces—run a sample essay through it and compare with a human tutor's notes. Check how often the repository is updated; the README mentions ongoing maintenance, but verify commit history. Also consider alternatives like generic prompt templates you might already have, or other specialized IELTS bots. Finally, assess whether the skill's output format fits your existing agent's workflow (e.g., markdown reports vs. inline comments).

Why it is worth tracking

IELTS Buddy is a focused, community-maintained project with 13 stars and a clear update mechanism via GitHub releases and OSS mirrors. It shows active thinking about making agent skills modular and reusable for a specific domain. If you're building a study routine around local AI, this is a practical starting point—and it's also a good example of how to structure skill collections for your own agent projects. For a quick overview, see the OpenAgentSkill listing. Re-evaluate as the repo grows and more users contribute feedback on skill quality.