# EduOrchestrate > Turn any AI coding agent into a daily learning command center — adaptive role plans, sourced research, daily emails. CLI-first. - **Type:** MCP server - **Install:** `agentstack add mcp-haroontrailblazer-eduorchestrate` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [haroontrailblazer](https://agentstack.voostack.com/s/haroontrailblazer) - **Installs:** 0 - **Category:** [Integrations](https://agentstack.voostack.com/c/integrations) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [haroontrailblazer](https://github.com/haroontrailblazer) - **Source:** https://github.com/haroontrailblazer/EduOrchestrate - **Website:** https://eolearn.vercel.app ## Install ```sh agentstack add mcp-haroontrailblazer-eduorchestrate ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # EduOrchestrate ### Turn any AI coding agent into a daily learning command center. **One command** turns a target job role into an adaptive plan, trend-aware research, daily emails, and a skill progression card — running locally inside the agent you already use. Your data, your machine. Quickstart · Features · Agents · CLI · How it works · FAQ --- ## Why EduOrchestrate? Most learning tools are SaaS dashboards that collect courses you never finish. EduOrchestrate is the opposite: a **vendor-neutral learning coach that lives in your CLI**. It turns "I want to become an Agentic AI Engineer" into a concrete daily loop — one core topic, one build artifact, one review — with sourced research and proof of progress. > **Job boards use AI to filter candidates. EduOrchestrate gives the learner AI to build a daily, sourced path to the job — one artifact at a time.** -  **Role → plan in one command.** Pick a target role and a first skill; get an adaptive plan with a fixed **30-day minimum**. -  **AI-native & agnostic.** The same `/eduorchestrate` skill works across Claude Code, Codex, Gemini, OpenClaw, Antigravity, Hermas, and any MCP agent. -  **Local-first & private.** Learning state stays in `data/`; SMTP secrets live in `.env.local` or GitHub Actions — never committed. -  **Zero dependencies.** A dependency-free Node CLI. Audit it in an afternoon. -  **Open source, MIT.** No account, no paywall, no telemetry. --- ## Quickstart From an agent terminal in your project: ```bash npx eduorchestrate ``` That's it. The interactive setup captures your profile (name, email, target role, current topic, **first-skill focus**, stage, plan length, daily email time, timezone, SMTP), writes the agent adapters, and generates your first 30-day plan. Requires Node 18+. Nothing to install globally — npx handles it. --- ## What it does FeatureDescription Adaptive role plansLearner-chosen length with a fixed 30-day minimum. One topic, one build artifact, one review loop per day. Trend-aware researchDaily digests of dated, sourced updates from official docs, YouTube search, and GitHub sample code — and only reshapes the roadmap when a trend improves employability. Daily email deliveryPreview with --dry-run, then send the day's plan over your own SMTP. Autonomous runnerA generated GitHub Actions workflow delivers your daily plan hands-free — no browser or agent session needs to stay open. Skill progression cardsPrebuilt, GitHub-style SVG terminal cards showing role, skill, momentum, and next action. Course source packsRole-specific official catalogs, free-course discovery links from major vendors, and newsletter feeds. Token-conscious harnessTells the agent what to load first, when to research, and what not to paste into chat. Progress & reviewsEvidence logging, weekly summaries, and a recommend-next that suggests your next skill after the plan window. ### Built-in role blueprints Agentic AI & LLM EngineerRAG, evals, agent workflows Full-Stack DeveloperWeb, cloud, CI/CD Cybersecurity AnalystSOC, SIEM, cloud security Data ScientistData eng, analytics, ML > For other roles, EduOrchestrate infers the closest blueprint and keeps the plan artifact-driven. --- ## Supported agents The same `/eduorchestrate` command is honored across runtimes — setup writes the right adapter for each. Where a runtime can't register slash commands, it treats `use EduOrchestrate` as equivalent to `/eduorchestrate`. | Agent | Adapter written on setup | |---|---| | **Claude Code** | `.claude/skills/eduorchestrate/SKILL.md` | | **Codex** / generic coding agents | `AGENTS.md` | | **Gemini** | `GEMINI.md` | | **OpenClaw** | `.eduorchestrate/adapters/openclaw/SKILL.md` | | **Antigravity** | `.eduorchestrate/adapters/antigravity/manifest.json` | | **Hermas** | `.eduorchestrate/adapters/hermas/manifest.json` | | **MCP-capable agents** | `.eduorchestrate/mcp/manifest.json` | --- ## The CLI Everything runs from one dependency-free CLI. State writes to `data/` and is git-ignored by default. ```bash # Plan npx eduorchestrate plan --role "Agentic AI and LLM Engineer" npx eduorchestrate plan --role "Agentic AI and LLM Engineer" --skill "RAG evaluation" --days 45 # Daily loop npx eduorchestrate harness # load the token-conscious harness npx eduorchestrate courses # role-specific course + newsletter packs npx eduorchestrate research --day 1 # dated, sourced trend digest npx eduorchestrate send-today --dry-run # preview the daily email npx eduorchestrate send-today # send for real (your SMTP) npx eduorchestrate log-progress --day 1 --completed "Built setup" --evidence "Repo link" # Review & cards npx eduorchestrate status # current role, skill, day, next action npx eduorchestrate weekly-summary --week 1 npx eduorchestrate progress-card # GitHub-style skill card (md + json) npx eduorchestrate terminal-card # terminal-style SVG card npx eduorchestrate recommend-next # next skill after the plan window # Setup helpers npx eduorchestrate setup-secrets # prints `gh secret set` commands for Actions email npx eduorchestrate doctor # verifies skill, adapters, config, env files ``` Command reference | Command | What it does | |---|---| | `plan` | Generate an adaptive plan (`--role`, `--skill`, `--days`; 30-day minimum enforced). | | `harness` | Write the compact execution harness (`data/eduorchestrate-harness.json` / `.md`). | | `courses` | Write role-specific learning sources (`data/learning-sources.json` / `.md`). | | `research` | Produce a dated, sourced trend digest for a given `--day`. | | `send-today` | Send today's plan email (`--dry-run` previews without sending). | | `log-progress` | Record completed work + evidence for a day. | | `status` | Print current role, skill, progress, current day, and next action. | | `weekly-summary` | Summarize a `--week` of progress. | | `progress-card` / `terminal-card` | Render learner-specific progression cards. | | `recommend-next` | Suggest the next skill once the plan window completes. | | `setup-secrets` | Print the `gh secret set` commands for Actions email delivery. | | `doctor` | Diagnose the repo's skill, adapter, config, and environment files. | --- ## How it works EduOrchestrate models the system as a **typed multi-agent workflow** — even when a single model executes it: ```mermaid flowchart LR R["Target role+ stage"] --> GM[GoalMapper] GM --> CP[CurriculumPlanner] CP --> RS[Researcher] RS --> BM[BuildMentor] BM --> FC[FocusCoach] FC --> EV[Evaluator] EV -->|next action| CP ``` | Agent | Responsibility | |---|---| | **GoalMapper** | Translate target role + stage into priority skills. | | **CurriculumPlanner** | Create milestones and a daily focus plan. | | **Researcher** | Gather dated, sourced updates from official docs and release notes. | | **BuildMentor** | Turn topics into practical projects with deliverables. | | **FocusCoach** | Limit scope and prevent resource overload. | | **Evaluator** | Review evidence and adjust the next action. | Each day becomes **one build action with visible proof** — research current references, practice one focused skill, preview or send the email, log evidence, keep the next action visible. --- ## Autonomous daily delivery EduOrchestrate doesn't need a browser window or an open agent session to deliver. On setup it writes `.github/workflows/eduorchestrate-daily.yml`, which runs in **GitHub Actions** at your configured time: - Runs `node bin/eduorchestrate.js send-today` on schedule. - Pulls SMTP credentials from **GitHub Actions secrets** (`setup-secrets` prints the exact `gh secret set` commands). - Advances the day from `schedule.startDate` and **stops after the plan window** instead of repeating the same day forever. > `send-today --dry-run` is always a safe preview. `send-today` sends real email. Adapters follow the same resume rule everywhere: on resume, load the harness/status, resolve today's day, and only send after SMTP setup is approved. --- ## Configuration & state | What | Where | Committed? | |---|---|---| | Plans, digests, logs, cards | `data/` | No — git-ignored | | SMTP secrets | `.env.local` or GitHub Actions secrets | No — never | | Schedule / profile config | `eduorchestrate.config.json` | Yes (no secrets) | --- ## Architecture ``` src/ ├── cli.js # dependency-free command router ├── planner.js # role blueprints + adaptive plan generation ├── adapters.js # writes per-agent adapter files ├── workflow.js # GitHub Actions runner generator ├── research.js # official docs / YouTube / GitHub-sample digests ├── courses.js # role-specific course + newsletter source packs ├── harness.js # token-conscious execution harness ├── progress.js # progress log, status, weekly summaries ├── progression-card.js # role/skill + terminal card generation ├── card-components.js # compact SVG component renderer ├── email.js # SMTP mailer └── config.js # config + data-dir resolution skills/eduorchestrate/ # canonical universal skill + assets + references tests/node/ # Node regression tests ``` --- ## Development ```bash git clone https://github.com/haroontrailblazer/EduOrchestrate.git cd EduOrchestrate npm test # run the Node regression suite npx . # run the CLI locally before publishing ``` The whole CLI is dependency-free and runs on Node 18+. There is no build step. --- ## FAQ Is it really free? What's the license? Yes — MIT-licensed and open source, with zero runtime dependencies. No account, no paywall, no telemetry. The only thing you pay for is whichever AI coding CLI you already use. Where does my data go? Local. Plans, research digests, progress logs, and cards write to data/, which is git-ignored. SMTP credentials live in .env.local or GitHub Actions secrets — never in committed config. Why a 30-day minimum? A month is the floor for building real, demonstrable momentum on a job role. You can request a longer plan, but if you ask for fewer than 30 days it still generates 30. Which agents does it work with? Claude Code, Codex and generic coding agents (AGENTS.md), Gemini, OpenClaw, Antigravity, Hermas, and any MCP-capable agent. See Supported agents. How does daily email work without keeping anything open? A generated GitHub Actions workflow runs send-today on schedule using your stored secrets. Run setup-secrets to print the commands that store SMTP credentials as Actions secrets. --- ## Contributing Contributions are welcome! Whether it's a new **role blueprint**, an **agent adapter**, better **course sources**, or docs — open an issue or PR. 1. Fork & branch from `main` 2. Keep it dependency-free and add/adjust tests in `tests/node/` 3. Run `npm test` 4. Open a PR with a clear description --- ## Star history --- ## License [MIT](LICENSE) © EduOrchestrate — your data, your machine. Built for learners who want proof, not passive curriculum. Star the repo if it helps you land the role. ## Source & license This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [haroontrailblazer](https://github.com/haroontrailblazer) - **Source:** [haroontrailblazer/EduOrchestrate](https://github.com/haroontrailblazer/EduOrchestrate) - **License:** MIT - **Homepage:** https://eolearn.vercel.app Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** yes - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/mcp-haroontrailblazer-eduorchestrate - Seller: https://agentstack.voostack.com/s/haroontrailblazer - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.