TEA Glossary

Terminology reference for Test Engineering Architect (TEA), starting with how its skills, agent, and CLIs are invoked.

Test Engineering Architect (TEA) Concepts

Invocation terms

Term	Definition
Skill	A packaged capability such as bmad-tea-trace. Start it in chat as /bmad-tea-trace (Claude Code, Cursor, Windsurf) or $bmad-tea-trace (Codex), or from a fresh session. TEA ships eight skills.
Module	An installable bundle of skills and shared resources. TEA is the tea module; the bmod-tea skill holds its settings record and knowledge base.
TEA agent	The bmad-tea persona. Load it to get a menu of codes such as TD and TR that start the skills. AT selects Automate red mode and CI selects Framework CI scope.
Chat command	A skill name or menu code typed into the assistant. The agent does the work, and you answer its questions.
Headless CLI	A tea-* terminal command such as tea-trace that starts an agent for one skill, then checks and publishes its outputs. Use it in CI or scripts. See Commands.

Concepts and skills

Term	Definition
ATDD	Acceptance Test-Driven Development. Automation red mode creates acceptance scaffolds before implementation and verifies intended missing-behavior failures. Request red mode through bmad-tea-automate.
Automation	The bmad-tea-automate skill. Red mode writes acceptance tests before implementation; expand adds coverage to existing code. Create executes generated tests and repairs test issues for up to three rounds, with product defects retained as findings.
Behavioral Evaluation Contract	The contract.json of an evaluation folder: the behaviors a target must show and the oracles that judge them. eval-quality compiles and seals it before a run.
Browser Automation	Playwright CLI and/or MCP servers enabling live browser interaction during test generation. CLI provides token-efficient stateless commands; MCP provides rich stateful automation. Configured via tea_browser_automation.
Burn-in Testing	Running tests multiple times (typically 5-10 iterations) to detect flakiness and intermittent failures.
CI Setup	The CI phase of bmad-tea-framework. Request CI scope through bmad-tea-framework.
Component Testing	Testing UI components in isolation using framework-specific tools (Cypress Component Testing or Vitest + React Testing Library).
Coverage Oracle	The list of things Trace must cover: formal requirements first, then contract or spec artifacts, resolvable external pointers, and journeys inferred from source. See Run Trace.
Coverage Traceability	Mapping coverage oracle items such as requirements, contract endpoints, external pointers, or inferred journeys to implemented tests with classification (FULL/PARTIAL/NONE) to identify gaps and measure completeness.
Engagement Model	One of five ways to use TEA: No TEA, TEA Solo, TEA Lite, TEA Integrated (Greenfield), or TEA Integrated (Brownfield). Enterprise adds a track to the Integrated models. See Engagement Models.
Epic-Level Test Design	Test planning per epic (Phase 4) focusing on risk assessment, priorities, and coverage strategy for that specific epic.
Evaluate	The bmad-tea-evaluate skill (menu code EV). Builds a scored evaluation, identifies gaps, and hands a CI plan to framework with CI scope. Uses the tea-evaluate runtime.
Evaluation Folder	The folder under the evaluations_folder setting that holds one evaluation: evaluation.json, the contract, the probes, the corpus, the committed baseline and the CI plan. Every tea-evaluate command takes it with --evaluation.
Fixture Architecture	Pattern of building pure functions first, then wrapping in framework-specific fixtures for testability, reusability, and composition.
Framework Setup	The bmad-tea-framework skill. Select framework only, CI only, or both by prompt; unclear CI scope in an interactive session prompts “Do you want CI too?” once. Create, Resume, Validate, and Edit select the operation.
Gate Decision	The derived release decision: PASS, CONCERNS, or FAIL. A human may grant WAIVED separately; trace validates and reports filed waivers while retaining its derived decision.
Knowledge Fragment	Individual markdown file in TEA’s knowledge base covering a specific testing pattern or practice (59 fragments total).
Network-First Pattern	Registering a network wait or intercept before the action that triggers the request, then waiting for its response.
NFR Evidence Audit	Validation of non-functional requirement evidence (security, performance, reliability, maintainability) against defined thresholds.
No TEA	Engagement model 1. Skip all TEA skills and keep the team’s existing testing approach. A valid choice when that approach already works.
Playwright Utils	The optional @seontechnologies/playwright-utils package of Playwright fixtures and helpers.
Risk-Based Testing	Testing approach where depth scales with business impact using probability × impact scoring (1-9 scale).
Run Key	The scope label in a per-scope output filename, such as epic-16, story-1-2, release-v1-2-0, or system. See Output Layout.
System-Level Test Design	Test planning at architecture level (Phase 3) focusing on testability review, ADR mapping, and test infrastructure needs.
tea-index.csv	Manifest file tracking all knowledge fragments, their descriptions, tags, tier, and fragment path.
TEA Academy	The teach-me-testing skill (menu code TMT): seven sessions with quizzes, saved progress, and examples matched to your role.
TEA Integrated	Engagement models 4 and 5. Full BMad Method integration with TEA skills across Phase 2, 3, 4, and the release gate. Greenfield starts from scratch; Brownfield adds a Phase 2 coverage baseline and a regression-hotspot focus.
TEA Lite	Engagement model 3. Use automate on existing features, with framework setup when needed.
TEA Solo	Engagement model 2. Standalone use of TEA skills without BMad Method integration; you bring your own requirements.
Test Priorities	Classification system for test importance: P0 (critical path), P1 (high value), P2 (medium value), P3 (low value).

BMad Method Terms

These terms come from the wider BMad Method and apply when TEA runs inside its planning lifecycle. The BMad Method documentation covers the rest of the ecosystem, and Game Dev Studio covers game-development vocabulary.

Core Concepts

Term	Definition
Agent	AI persona with a role, such as analyst, architect, developer, or test architect.
BMad	Breakthrough Method of Agile AI-Driven Development. An AI-assisted development framework with agents and guided workflows.
BMad Method	AI-assisted development method covering planning, architecture, implementation, and testing.
BMM	BMad Method Module. The development agents and workflows supplied by BMad Method.
Scale-Adaptive System	Workflow selection that adjusts planning depth to the project.
Workflow	A guided process that uses an AI agent to produce an artifact or complete a task.

Scale and Complexity

Term	Definition
BMad Method Track	Full product planning track using PRD + Architecture + UX. Best for products, platforms, and complex features.
Enterprise Method Track	Extended planning track adding Security Architecture, DevOps Strategy, and Test Strategy. Best for compliance needs and multi-tenant systems.
Planning Track	Planning path selected for the scope and complexity of the work.
Quick Flow Track	Fast implementation track using tech-spec only. Best for bug fixes, small features, and clear-scope changes.

Planning Documents

Term	Definition
Architecture Document	BMad Method/Enterprise. System-wide design document defining structure, components, data models, integration patterns, security, and deployment.
Epics	Groups of related stories that deliver a feature.
PRD	BMad Method/Enterprise. Product Requirements Document containing vision, goals, FRs, NFRs, and success criteria. Focuses on WHAT to build.
Product Brief	Phase 1. Optional strategic document capturing product vision, market context, and high-level requirements before detailed planning.
Tech-Spec	Quick Flow. Technical plan covering the problem, solution, file changes, and testing.

Workflow and Phases

Term	Definition
Phase 0: Documentation	Brownfield. Conditional prerequisite phase creating codebase documentation before planning. Only required if existing docs are insufficient.
Phase 1: Analysis	Discovery phase including brainstorming, research, and product brief creation. Optional for Quick Flow, recommended for BMad Method.
Phase 2: Planning	Required phase creating formal requirements. Routes to tech-spec (Quick Flow) or PRD (BMad Method/Enterprise).
Phase 3: Solutioning	BMad Method/Enterprise. Architecture design phase including creation, validation, and gate checks.
Phase 4: Implementation	Required sprint-based development through story-by-story iteration using sprint-planning, bmad-build (create-story and dev-story before BMAD 6.12), and code-review workflows.
Quick Spec Flow	Fast-track workflow for Quick Flow projects going straight from idea to tech-spec to implementation.
Workflow Init	Initialization workflow creating bmm-workflow-status.yaml, detecting project type, and determining planning track.
Workflow Status	Universal entry point checking for existing status file, displaying progress, and recommending next action.

Agents and Roles

Term	Definition
Analyst	Agent that initializes workflows, conducts research, creates product briefs, and tracks progress. Often the entry point for new projects.
Architect	Agent designing system architecture, creating architecture documents, and validating designs. Primary agent for Phase 3.
BMad Master	BMad Core agent that runs party mode and guides work across modules.
DEV	Developer agent implementing stories, writing code, running tests, and performing code reviews. Primary implementer in Phase 4.
Party Mode	Multi-agent collaboration feature where agents discuss challenges together. BMad Master orchestrates, selecting 2-3 relevant agents per message.
PM	Product Manager agent creating PRDs and tech-specs. Primary agent for Phase 2 planning.
SM	Scrum Master agent managing sprints, creating stories, and coordinating implementation. Primary orchestrator for Phase 4.
TEA	Test Engineering Architect. Agent responsible for test strategy, quality gates, NFR planning, and NFR evidence audit. Spans Phase 3, Phase 4, and the release gate.
Technical Writer	Agent specialized in creating technical documentation, diagrams, and maintaining documentation standards.
UX Designer	Agent creating UX design documents, interaction patterns, and visual specifications for UI-heavy projects.

Status and Tracking

Term	Definition
bmm-workflow-status.yaml	Phases 1-3. Tracking file showing current phase, completed workflows, and next recommended actions.
DoD	Definition of Done. Criteria for completing work, such as passing tests, reviewed code, and updated docs.
Epic Status Progression	backlog → in-progress → done.
Gate Check	An implementation-readiness review of the PRD, architecture, and epics before Phase 4.
Retrospective	Workflow after each epic capturing learnings and improvements for continuous improvement.
sprint-status.yaml	Phase 4. Single source of truth for implementation tracking containing all epics, stories, and their statuses.
Story Status Progression	backlog → ready-for-dev → in-progress → review → done.

Project Types

Term	Definition
Brownfield	Existing project with established codebase and patterns. Requires understanding existing architecture and planning integration.
Convention Detection	Quick Flow. Feature auto-detecting existing code style, naming conventions, and frameworks from brownfield codebases.
document-project	Brownfield. Workflow analyzing and documenting existing codebase with three scan levels: quick, deep, exhaustive.
Feature Flags	Brownfield. Implementation technique for gradual rollout, easy rollback, and A/B testing of new functionality.
Greenfield	New project starting from scratch with freedom to establish patterns, choose stack, and design from clean slate.
Integration Points	Brownfield. Specific locations where new code connects with existing systems. Must be documented in tech-specs.

Implementation Terms

Term	Definition
Context Engineering	Loading the standards and project facts an AI agent needs for its task.
Correct Course	Workflow that assesses a change during implementation and proposes updates to the plan.
Shard / Sharding	Splitting large planning documents into section-based files for LLM optimization. Phase 4 workflows load only needed sections.
Sprint	Time-boxed period of development work, typically 1-2 weeks.
Sprint Planning	Workflow initializing Phase 4 by creating sprint-status.yaml and extracting epics/stories from planning docs.
Story	A unit of implementable work with acceptance criteria. Related stories form an epic.
Story Context	Implementation guidance embedded in the story spec that bmad-build plans (the create-story story file before BMAD 6.12), referencing existing patterns and approaches.
Story File	Markdown file containing story description, acceptance criteria, technical notes, and testing requirements.
Track Selection	Automatic analysis by bmad suggesting appropriate track based on complexity indicators. User can override.

See Also

• TEA Overview: the eight skills and the phase lifecycle
• Evaluate Your First Skill: the Evaluate terms above in a first scored run
• Engagement Models: the five models defined above
• TEA Knowledge Base: fragment index
• TEA Command Reference: skill reference
• TEA Configuration: config options