Visual Maps
Visual index for Datarim maps. Load this entry first, then only the diagram fragment needed for routing, stage flow, or dependency orientation.
Overview
Visual Maps provides Mermaid workflow diagrams for navigation and orientation within the Datarim pipeline. It visualizes pipeline routing by complexity level, stage process flows, command-agent relationships, agent-skill dependencies, and Consilium panel compositions. Loaded on demand when an agent needs to understand routing or component relationships.
Fragment Routing
The skill is split into 5 focused fragments: pipeline-routing (complexity-based routing), stage-process-flows (/dr-init through /dr-archive), content-and-management-flows (/dr-write, /dr-edit, /dr-dream, /dr-optimize, /dr-addskill), utility-and-dependencies (command-agent-skill relationships), and panels-and-quality (Consilium panels and quality rule mapping).
Pipeline Routing
A visual decision tree showing how task complexity (L1-L4) determines the pipeline route. Level 1 tasks follow the shortest path (init, do, reflect, archive). Level 4 tasks traverse the full pipeline including PRD, design, phased implementation, QA, and compliance stages. Optional stages are marked with brackets.
Stage Process Flows
Each pipeline stage has a visual process flow: /dr-init reads the task and routes by level; /dr-prd runs discovery, generates approaches, and consults the user; /dr-plan breaks down components with a strategist gate for L3-4; /dr-do implements iteratively with pre-flight checks; /dr-qa runs four verification layers; /dr-compliance detects task type and applies type-specific checklists.
Component Relationships
- Command-Agent map — which agent handles each
/dr-command, including secondary agents for L3-4 complexity - Agent-Skill dependencies — which skills each of the 14 agents loads (e.g., architect loads datarim-system, tech-stack, performance, security, consilium)
- Consilium panels — preset agent combinations for architecture, code design, production readiness, content, and knowledge discussions
Quality Rules by Stage
Maps the 15 AI Quality rules to specific pipeline stages. For example, /dr-plan applies rules #1 (stubbing), #5 (DoD), #6 (corner cases), #7 (skeleton), #11 (boundaries). /dr-do applies #2 (TDD), #3 (method size), #8 (iterative), #9 (cognitive load).
When It's Used
Loaded on demand for navigation and orientation. Not loaded for simple tasks where the next step is obvious. Primarily useful for onboarding, debugging pipeline flow, and documenting the framework.