| name | exeris-platform-task-classifier |
| description | Triage step for any incoming exeris-platform task. Use FIRST — before routing or implementation — to classify a Studio / LSP / studio-backend request into task type (architecture / backend / LSP / frontend / docs / cross-build), scope, severity, and primary contract risk, and to name the specialist agent that should own it. Invoke whenever the owning module or the main risk is not yet obvious, or when scope may cross modules. |
Exeris Platform Task Classifier
Purpose
Classify incoming work before execution starts. Triage only — no implementation.
Output Contract
Return exactly:
task_class (ARCHITECTURE | BACKEND_IMPLEMENTATION | LSP_PROTOCOL | FRONTEND_IMPLEMENTATION | DOCS_ADR | CROSS_BUILD | MULTI_DOMAIN)
scope (single-module | cross-module | cross-build [Maven↔npm])
severity (low | medium | high | critical)
primary_risk
recommended_primary_agent
Classification Heuristics
ARCHITECTURE: module placement, no-parallel-metamodel, open-core boundary, LSP-vs-backend-HTTP scope.
BACKEND_IMPLEMENTATION: exeris-studio-backend Java workspace-state code.
LSP_PROTOCOL: exeris-platform-lsp server, wire surface, exeris/* methods, idempotent write-back.
FRONTEND_IMPLEMENTATION: exeris-studio-frontend Angular + embedded React UI.
DOCS_ADR: README target architecture, ROADMAP milestones, cross-repo ADR registry.
CROSS_BUILD: changes require coordination across Maven reactor and npm package.
MULTI_DOMAIN: at least two classes above are first-order concerns.
Guardrails
- Preserve no-parallel-metamodel discipline.
- Preserve LSP as the wire boundary for domain shape.
- Preserve open-core boundary.
- Preserve
exeris/* namespace for custom LSP methods.
- If uncertain between two classes, emit
MULTI_DOMAIN and state both dominant concerns.
Completion Criteria
All five output fields present and each justified in 1-2 concise bullets.