Skip to content

v0.3.0 - Learning Path Support, Engineering Bootcamp and New Courses for Beginners

Latest
Latest

Choose a tag to compare

View all tags
@ishands ishands released this 29 Apr 15:28
v0.3.0
06507f4
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

v0.3.0 — 2026-04-29

This is the largest release since open-sourcing. It adds seven new courses, the first curated learning path, and a round of improvements to the first-run experience surfaced by beta testing.

Learning paths

Upstack now supports learning paths — a sequence of courses bundled around a single project you pick at enrolment. Each course adds a component to your project through an integration task. The capstone at the end is assembly, not a new course.

The first curated path is the Engineering Bootcamp: seven fundamentals courses for fresh-graduate or early-career software engineers, with three capstone project options (Personal Library Catalogue, Recipe Manager, Job Application Tracker). Use /start-learning-path to enrol.

Seven new courses

Seven standalone engineering fundamentals courses, each usable independently or as part of the Engineering Bootcamp:

  • Git Fundamentals — local repo, branching, pull requests
  • Markdown Fundamentals — READMEs, PR descriptions, technical documentation
  • OOP Fundamentals — classes through to SOLID principles and design patterns
  • DSA Fundamentals — arrays, hash maps, recursion, Big O intuition
  • Code Quality Fundamentals — naming, code smells, refactoring
  • Testing Fundamentals — test scenario identification, test case design, acceptance criteria
  • Agile Fundamentals — Scrum roles, ceremonies, sprint artefacts

Skills now work as slash commands

Skills have moved to the cross-provider .agents/skills/ location (the agentskills.io standard). As a side effect, /start-course, /configure-profile, /check-progress, and the rest now work as native slash commands in Claude Code — this was broken in the previous layout.

First-run improvements

  • The tutor now follows a clearer decision tree at session start — profile check, then active path/course check, then calibration. Several gaps in the first-run flow are closed.
  • configure-profile accepts a pasted CV or professional bio as a fast path — the tutor parses what it can and only asks follow-up questions for what's missing.
  • Skill levels are now presented as a plain 1–5 scale with everyday anchor phrases. No framework jargon in learner-facing language.

Tested with

Claude Code Sonnet 4.6 (High effort), Gemini 3.1 Pro via Antigravity (High effort), and GitHub Copilot Free (Auto). Free-tier performance may vary — see AI Tutor Limitations for details.

Assets 2
Loading