Architecture Overview
System layers
Section titled “System layers”The Scoutica Protocol is organized into four layers. Three are live today; the Trust Layer is on the roadmap.
| Layer | Status | Components |
|---|---|---|
| Candidate Layer | ✅ Live | Documents, Scoutica CLI, Skill Card, configured AI provider |
| Distribution Layer | ✅ Live | GitHub, Registry, scoutica.json |
| Consumption Layer | ✅ Live | AI Agents, Job Boards, ATS Systems, Widgets |
| Trust Layer | 🔜 Roadmap | Blockchain, Soulbound Tokens, Zero-Knowledge Proofs |
Candidate Layer
Section titled “Candidate Layer”The Candidate Layer is where Skill Cards are created. Candidates use the Scoutica CLI or any AI assistant to generate a structured set of files (profile.json, rules.yaml, evidence.json, scoutica.json) from their documents. Extraction runs locally; a remote-capable provider may receive the full generated prompt and document text.
Distribution Layer
Section titled “Distribution Layer”Once created, a Skill Card is published to a GitHub repository. The scoutica.json discovery file at the repo root makes the card machine-readable. Cards can also be registered in the central registry index for broader discovery.
Consumption Layer
Section titled “Consumption Layer”AI agents, job boards, ATS systems, and embeddable widgets all consume cards from the Distribution Layer. Consumers fetch card files directly from GitHub over HTTPS and evaluate candidates locally using their own scoring logic.
Trust Layer (future)
Section titled “Trust Layer (future)”The Trust Layer will add cryptographic verification via Soulbound Tokens on Base L2 and Zero-Knowledge Proofs. This enables Verification Level 4 (blockchain-verified) in the trust model.
Two-part architecture
Section titled “Two-part architecture”The project has a strict separation between two concerns:
| Part | What | Visibility |
|---|---|---|
| The Protocol | Open standard, schemas, agent rules, registry API spec, CLI tools, smart contracts, templates | Public — anyone can clone, fork, build on it |
| Your Skill Card | Your personal profile, Rules of Engagement, evidence, salary floor | Private — never committed to the protocol repo |
The protocol is the network. Your skill card is your node on it.
When you clone this repo, you get:
- JSON Schemas to validate your card
- Agent rule templates to evaluate opportunities
- Registry API spec to run your own node
- CLI tools to publish and discover cards
You do not get anyone’s private data. Each card lives in its owner’s private directory.
Card generation flow
Section titled “Card generation flow”Cards flow from source documents through the CLI to GitHub:
Documents → scoutica scan → configured AI provider → Skill Card files → GitHub repo → Registry / DiscoveryFrom GitHub, any consumer can resolve a card by fetching the four canonical files:
scoutica.json ← discovery entry pointprofile.json ← skills, experience, seniorityrules.yaml ← engagement rules, salary floor, auto-reject criteriaevidence.json ← GitHub repos, certifications, portfolio links