Get Started
In this section
Back to use cases

Educator

ℹ Illustrative scenario — a composite archetype, not a real named adopter.

Meet Prof. Sarah Kim

Sarah teaches a graduate course on AI-augmented knowledge work at a research university. Her students come from diverse backgrounds — computer science, biology, business, humanities. She needs a framework that’s rigorous enough for the CS students and accessible enough for the humanities students.

The Challenge

Teaching “how to work with AI agents” is hard because there’s no standard curriculum. Every tool has its own approach. Students learn prompting techniques but not knowledge architecture — they can ask an AI a good question but can’t organize their project so the AI gives consistently good answers. Sarah wants to teach the structural layer, not just the interaction layer.

How aDNA Helps

Sarah uses aDNA itself as the teaching tool. Students don’t just read about knowledge architecture — they build one. The course vault is both the syllabus and the subject.

Week 1-3: Students navigate this vault (aDNA.aDNA/), reading concept files and following the beginner tutorials. They learn the triad, governance files, and the question test by seeing them in action.

Week 4-6: Students fork the base template and create their own project vault. They extend the ontology for their thesis domain, write context files, and experience how structure improves agent output.

Week 7-9: Students run a mini-campaign — 3 missions, each with deliverables. They practice mission decomposition, session tracking, and AARs.

Week 10: Students present their vaults. The dual-audience principle is the grading rubric — a non-expert should be able to navigate the vault and understand the domain.

What Their Vault Looks Like

# Course vault (shared reference):
ai_literacy_course.aDNA/
├── what/
│   ├── context/         # Readings, research summaries
│   ├── concepts/        # Student-authored concept files
│   └── assignments/     # [EXT] Assignment specifications
├── how/
│   ├── campaigns/       # Course syllabus as a campaign
│   ├── missions/        # Weekly assignments as missions
│   ├── sessions/        # Student work logs
│   └── templates/       # Assignment templates
└── who/
    ├── governance/      # Course policies, grading rubric
    └── team/            # Student roster, office hours

# Each student also has:
student_thesis.aDNA/     # Individual project vault

Outcome

Students leave the course able to structure knowledge for AI agents — not just prompt them. The dual-audience principle improves their writing across all contexts. Several students adopt aDNA for their thesis work. The course vault itself becomes a growing knowledge resource maintained across cohorts.

Typical Ontology Extensions

EntityTriadPurpose
assignmenthow/Assignment specifications with rubrics and deadlines
student_projectwhat/Registry of student vault forks with progress tracking
readingwhat/Course readings with discussion prompts

The teaching kit

A 3-week curriculum: nine tutorials in pedagogical order, each self-contained and classroom-sized.

Week 1 — orientation

  1. Navigate an aDNA Vault — 15 min · Beginner — guided tour of a live vault. Students learn the triad and governance files by walking them.
  2. Apply the Question Test — 15 min · Beginner — sort 10 sample items into what/how/who. First structural-thinking exercise.
  3. Create Your First CLAUDE.md — 20 min · Beginner — students write the agent-orientation file for their forked project vault.

Week 2 — authoring

  1. Write a Context File — 30 min · Intermediate — a quality-rubric-scored context file ready for their vault library.
  2. Extend the Ontology — 25 min · Intermediate — a domain-specific entity type with directory, AGENTS.md, and template.
  3. Design a Mission — 25 min · Intermediate — decompose multi-session work into claimable objectives.

Week 3 — composition

  1. Build a Lattice — 30 min · Advanced — a validated .lattice.yaml as a directed graph of modules.
  2. Run a Campaign — 30 min · Advanced — a phased multi-mission initiative with quality gates.
  3. Federate a Vault — 30 min · Advanced — export, import, and compose lattices across aDNA instances.