# AIOSON Glossary > **Who this is for:** anyone reading the other docs and tripping over jargon. > **Reading time:** reference — read what you need. > **How to use:** Ctrl+F is your friend. Terms in alphabetical order. Each entry has a **short definition** + **concrete example** + **where to learn more**. --- ## Agent **Definition:** a virtual specialist with its own prompt, its own rules, and a specific responsibility. Lives in `.aioson/agents/.md`. **Example:** `@dev` is the agent that implements code. `@qa` is the one that writes tests. `@pentester` is the one that tries to break the feature on purpose. **How to invoke:** type `@name` in your AI client (Claude, Codex e OpenCode) or describe the intent in natural language ("let's implement the feature"). --- ## AIOSON **Definition:** the framework itself. An npm package (`@jaimevalasek/aioson`) that installs a structure of agents, rules, and tools in your project. **Example:** running `npx @jaimevalasek/aioson init my-app` creates the `.aioson/` folder in your project. --- ## AIOSON Cloud / aioson.com **Definition:** the hosted service for publishing and installing squads, genomes, and skills across projects. **Example:** you create a squad of agents for the legal domain in your project, run `aioson system:publish`, and any other project can run `aioson system:install `. **Where to learn more:** `5-reference/aioson-com-store.md`. --- ## Artifact **Definition:** any file that an agent creates or updates as output of its work. **Examples:** - `.aioson/context/project.context.md` — project context (created by `@setup`) - `.aioson/context/features//spec.md` — feature spec (created by `@product`/`@analyst`) - `.aioson/context/dossier//` — feature dossier (created by the agent chain) **Why it matters:** Article III of the Constitution says important work leaves an artifact. Without an artifact, the work "didn't happen" officially. --- ## Brains **Definition:** procedural memory of an agent — structured knowledge in "nodes" (Zettelkasten style) that the agent loads on demand. **Example:** `@site-forge` has brains with 14 nodes about how `pt.squarespace.com` structures its pages, used when you ask for a clone. **Where it lives:** `.aioson/brains//`. --- ## Briefing **Definition:** a pre-PRD document that frames a problem before it becomes a feature. Output of the `@briefing` agent. **Example:** you have 5 scattered notes about an idea. `@briefing` transforms them into a structured briefing with problem, hypotheses, and discovery frames. --- ## Classification (MICRO / SMALL / MEDIUM) **Definition:** the size/risk profile calculated from 3 factors (user types, external integrations, business rules). It determines the depth, work budget, and review coverage—not the stage chain. **How it works:** - 0–1 points → **MICRO**: concise PRD/plan and focused QA - 2–3 points → **SMALL**: normal feature depth and focused regression - 4–6 points → **MEDIUM**: deeper named-risk, integration, and negative-path evidence All three use Product → Sheldon → Planner → DEV → QA after setup. **Where it appears:** `classification:` in the `project.context.md` frontmatter. --- ## AI Client **Definition:** the program where you actually talk to the AI. AIOSON is client-agnostic — it works with several. **Officially supported:** Claude Code, Codex CLI, OpenCode. **Also works via terminal:** VS Code, Cursor, Windsurf, JetBrains IDEs, Zed, Antigravity (any that opens a terminal). **Where to learn more:** `5-reference/clientes-ai.md`. --- ## Committer **Definition:** the `@committer` agent that generates professional commit messages from the diff and context. **Example:** after implementing a feature, `@committer` reads what changed, proposes the message, shows a checkbox in the terminal, and commits. --- ## Constitution **Definition:** the 6 (currently 7, with Zero Trust) principles in `.aioson/constitution.md`. No agent can contradict an article. **Full detail:** [Why it exists](./why-it-exists.md#the-6-principles-of-the-constitution). --- ## Context Pack **Definition:** a minimal, curated set of files that an agent reads to do its work — instead of scanning the entire project. **Why it matters:** saves context window. `@dev` for a "add a form field" feature doesn't need to read the whole project. --- ## Continuity (agent-chain-continuity) **Definition:** the system that ensures if a session drops in the middle of a feature, the next session can resume without losing context. Implemented in phases (1 to 8) across Mar–May 2026. **Main components:** `dossier`, `dev-state.md`, `dev-resume`, `handoff-protocol.json` v2, drift detection. --- ## Design Skill **Definition:** a skill specialized in a visual system (colors, typography, spacing, components). Can be a native skill or a hybrid created by `@design-hybrid-forge`. **Ready-made examples:** Clean SaaS UI, Aurora Command UI, Cognitive Core UI, Bold Editorial UI, Warm Craft UI, Glassmorphism UI, Neo Brutalist UI. **Where to choose:** during `aioson init`, in the wizard, or later with `@setup`/`@ux-ui`. The recommended route is the `interface-design` engine driven by your own **reference images**, extracted once into an `identity.md` the engine applies; the presets remain an explicit alternative. --- ## Dossier (Feature Dossier) **Definition:** a folder `.aioson/context/dossier//` that contains **everything** about a feature in progress: spec, plan, decisions, touched code, research index, status. **Why it matters:** it's the single point any agent entering the feature consults. Replaces "I'll re-read the entire history." **Commands:** `aioson dossier:init`, `dossier:show`, `dossier:add-research`, `dossier:audit`. --- ## Genome **Definition:** "cognitive DNA" of a persona — a structured YAML with personality traits (DISC, Enneagram, Big Five, MBTI, HEXACO-H), style, tone, and advisor instructions. **What it's for:** creating advisors (personalized assistants) that reason like a specific person. Example: generate a "Steve Jobs style" advisor to review pitch decks. **Current version:** Genome 4.0 with `anchor_prompt`, `relations`, `hexaco_h`, `trait_interactions`. **Where to learn more:** `5-reference/genome-4.0-spec.md` (in progress). --- ## Handoff **Definition:** the moment when an agent finishes its part and passes to the next, leaving the artifacts ready. **Where it's recorded:** `.aioson/context/handoff-protocol.json` (v2). --- ## Live Session **Definition:** a "tracked" session — an envelope that records each milestone in the local SQLite for the dashboard to display. **Commands:** `aioson live:start`, `live:status`, `live:handoff`, `live:close`. --- ## Memory **Definition:** AIOSON's persistent memory. Recently gained an **active retrieval** layer — agents can search for relevant memories, not just read everything. **Where it lives:** `.aioson/memory/` (in the project) and `~/.claude/projects//memory/` (per user, for Claude Code). --- ## Neo **Definition:** the `@neo` agent — intelligent router. It looks at the project state and tells you which agent makes sense to invoke now. **When to use:** when you don't know how to continue ("show project status", "where do I start?"). Recently gained a full ecosystem catalog and *ecosystem-inquiry* mode. --- ## Pentester **Definition:** `@pentester` agent that does adversarial security review — scans OWASP Top 10, LLM Top 10, supply chain, and maps attack surfaces. **When to use:** when the user explicitly requests an adversarial security review or an approved plan/QA finding names that need. It is disabled by default at every classification. --- ## Pipeline / Workflow **Definition:** the canonical tracked-feature sequence: optional raw-source Briefing → Briefing Refiner → user approval, then Product → Sheldon → Planner → DEV → QA. Classification changes depth and budget, not the sequence. Analyst, Architect, Discovery Design Doc, PM, UX/UI, and Scope Check are explicit consultants; Tester, Pentester, and Validator are opt-in reviewers. **Central command:** `aioson workflow:next .` — shows which agent is next. --- ## Plan **Definition:** either source notes in `/plans/` (project root), or the canonical `implementation-plan-{slug}.md` owned by `@planner`. **What they're for:** source notes may seed Product/Briefing. The canonical implementation plan maps PRD capabilities and ACs to vertical stages, exact files, checks, and production-path evidence. --- ## Profiler (researcher / enricher / forge) **Definition:** the 3-stage pipeline that creates a Genome from a public figure. 1. `@profiler-researcher` — collects raw material. 2. `@profiler-enricher` — analyzes cognitively. 3. `@profiler-forge` — generates the Genome 4.0 and the advisor. --- ## Project Pulse **Definition:** the file `.aioson/context/project-pulse.md` — the live global state of the project. Read at the start of each session, updated at the end. **What it's for:** give any incoming agent a snapshot of "what's happening in this project right now." --- ## Rules **Definition:** *hard* project rules, in `.aioson/rules/*.md`. Unlike loose instructions in a prompt, rules are read automatically by all applicable agents. **Example:** `security-baseline.md` declares the `SEC-SBD-01..08` controls that `@dev` and `@qa` consume. --- ## Skill **Definition:** a pluggable instruction package. Types: - **Process skills** — methodologies (e.g., `aioson-spec-driven`) - **Design skills** — visual systems (e.g., `clean-saas-ui`) - **Static skills** — fixed domain knowledge - **Dynamic skills** — generate instructions at runtime **Where they live:** `.aioson/skills/` in the project, or globally. --- ## Squad **Definition:** a group of custom agents — tailor-made for a domain the standard team doesn't cover well. **Example:** a "legal compliance" squad with agents `@regulator`, `@attorney`, `@auditor`, under the command of `@squad`. **Commands:** `aioson squad:scaffold`, `squad:agent-create`, `squad:doctor`. --- ## Tester **Definition:** opt-in `@tester` agent — systematic test engineering for already-implemented apps. Enable it explicitly when the proportional QA review is intentionally not enough (legacy, brownfield, gaps across 3+ modules). **Recent:** gained *coverage quality tier* and *test smell audit*. --- ## Validator **Definition:** opt-in `@validator` agent — technically validates a feature against an enabled success contract in a fresh context. It is never activated by classification alone. --- ## Workflow See **Pipeline / Workflow**. --- ## Zero Trust **Definition:** Article VII of the Constitution. Security is not an "optional feature" — it's the default state. Every feature goes through automatic baseline checks. --- Term not found? Search the [Ecosystem map](./ecosystem-map.md) or the [agents guide](../README.md).