A local-first, dependency-free CLI that turns any AI coding runtime into a daily learning orchestrator. A single npx command profiles a target job role, maps it to the smallest useful skill set, generates an adaptive 30-day plan, and ships daily research and build tasks by email — converting learning into portfolio-ready evidence, not passive consumption.
EduOrchestrate is a universal agent skill and daily learning orchestrator for job-role learners. A single npx command profiles the learner, generates an adaptive 30-day-minimum plan, wires up daily email delivery and GitHub Actions automation, and runs entirely locally across seven AI runtimes — not a motivational chatbot, but a training orchestrator that protects focus, resolves live sources, and converts learning into job-role evidence.
A walkthrough inside a coding agent — profiling a job role, generating the 30-day plan, resolving live sources, and rendering progress cards from logged evidence.
How a single day runs end to end — from loading the harness to rendering proof of work, across the agent network.
npx eduorchestrate harness loads the agent's execution contracts so the runtime knows exactly how to act.
status reports the active plan day, the next action, and current momentum.
research --day N builds a dated digest, resolving the top YouTube video and GitHub repo for the topic.
links --day N resolves concrete results into data/links.json for reuse.
Execute the day's practical task against a proof target and a set of review questions.
log-progress captures completion, evidence links, duration, confidence and blockers.
progress-card and terminal-card generate the card UI from logged evidence — no hand-authoring.
The capabilities that make EduOrchestrate a focus-protecting training orchestrator, not a motivational chatbot.
Research digests resolve concrete, shareable links — the current top YouTube video and the top GitHub repo by stars — with a soft fallback to search.
Enforces a 30-day minimum and caps each day at one core topic, one build artifact and one review loop.
Everything runs locally — plans, digests, logs and cards are written to data/ in your repo, with an optional offline mode.
Converts learning into portfolio-ready artifacts. Passive consumption does not count as completion.
A GitHub Actions runner ships the daily plan by email at a configured time, fully autonomously.
Runs inside Claude Code, Codex, Gemini and any MCP-capable coding agent — no lock-in.
Six typed agents, each owning one rule of the learning loop — coordinated into a single daily workflow.
Translates the target role and current stage into a prioritized skill set — the smallest set that produces role-relevant artifacts in 30 days.
Builds milestones and daily focus plans: less theory when weekly hours are low; build work every day.
Gathers dated, sourced updates from official docs and release notes — and never invents fetched results.
Turns each topic into a practical project so every guide leads to a portfolio-ready artifact.
Caps scope to prevent overload — one topic and one artifact per day.
Reviews logged evidence and adjusts the next move — advance, repeat or narrow.
