Features
Features
{"content":"Valdr features are the product capabilities you use to turn ad-hoc agent runs into a repeatable operating system for software work. These pages explain what each feature does, why it matters, and how it fits with the UI and MCP reference docs.\n\nStart here when you want the outcome and workflow view. Use the UI and MCP sections when you need screen-by-screen or tool-by-tool details.\n\nIn these docs, a **Valdr Workflow** is a saved, versioned process with durable runs. **Orchestration** coordinates specialized agents, while **Team Capabilities** provide the reusable instructions those agents follow.\n\n{{\u003c cards \u003e}}\n {{\u003c card link=\"knowledge/\" title=\"Workspace Knowledge\" subtitle=\"Persistent workspace context, cross-project code search, and Agent Memory Notebooks\" icon=\"server\" \u003e}}\n {{\u003c card link=\"agent-sessions/\" title=\"Agent Sessions\" subtitle=\"Launch, inspect, message, and chain agent execution through MCP\" icon=\"terminal\" \u003e}}\n {{\u003c card link=\"workflows/\" title=\"Durable Workflows\" subtitle=\"Compose agents, tools, review evidence, and human gates into versioned processes that survive restarts\" icon=\"arrows-expand\" \u003e}}\n {{\u003c card link=\"team-capabilities/\" title=\"Team Capabilities\" subtitle=\"Reusable operational behavior that teaches agents how your team works\" icon=\"user-group\" \u003e}}\n {{\u003c card link=\"orchestration/\" title=\"Multi-Agent Orchestration\" subtitle=\"Coordinate planners, executors, reviewers, and auditors through explicit workflow control\" icon=\"switch-horizontal\" \u003e}}\n {{\u003c card link=\"reviewable-workflows/\" title=\"Reviewable AI Workflows\" subtitle=\"Evidence-backed review, audit, scoring, and approval gates for agent work\" icon=\"check-circle\" \u003e}}\n{{\u003c /cards \u003e}}\n\n## Platform stack\n\nThe feature pages build upward through the Valdr platform model:\n\n| Layer | Feature | What changes |\n|-------|---------|--------------|\n| **Context** | [Workspace Knowledge](knowledge/) | Agents start with persistent engineering context instead of pasted summaries |\n| **Runtime** | [Agent Sessions](agent-sessions/) | Agent work becomes persistent, inspectable orchestration runtime state |\n| **Process** | [Durable Workflows](workflows/) | Repeated agent processes become saved, resumable workflows instead of terminal choreography |\n| **Operational behavior** | [Team Capabilities](team-capabilities/) | Team standards become reusable agent behavior instead of repeated prompts |\n| **Coordination** | [Multi-Agent Orchestration](orchestration/) | Specialized agents hand work off through explicit workflow control |\n| **Governance** | [Reviewable AI Workflows](reviewable-workflows/) | Agent output gets evidence-backed review, scoring, and approval gates |\n\n## Start with workflows\n\nSee [Durable Workflows](workflows/) for the product value, then [run your first workflow](/valdr/docs/ui/workflows/first-workflow/) without writing YAML.","description":"Product-level guides to the Valdr features that make agent work more reliable, inspectable, and reusable.","title":"Features"}
Valdr features are the product capabilities you use to turn ad-hoc agent runs into a repeatable operating system for software work. These pages explain what each feature does, why it matters, and how it fits with the UI and MCP reference docs.
Start here when you want the outcome and workflow view. Use the UI and MCP sections when you need screen-by-screen or tool-by-tool details.
In these docs, a Valdr Workflow is a saved, versioned process with durable runs. Orchestration coordinates specialized agents, while Team Capabilities provide the reusable instructions those agents follow.
Persistent workspace context, cross-project code search, and Agent Memory Notebooks
Launch, inspect, message, and chain agent execution through MCP
Compose agents, tools, review evidence, and human gates into versioned processes that survive restarts
Reusable operational behavior that teaches agents how your team works
Coordinate planners, executors, reviewers, and auditors through explicit workflow control
Evidence-backed review, audit, scoring, and approval gates for agent work
Platform stack
The feature pages build upward through the Valdr platform model:
| Layer | Feature | What changes |
|---|---|---|
| Context | Workspace Knowledge | Agents start with persistent engineering context instead of pasted summaries |
| Runtime | Agent Sessions | Agent work becomes persistent, inspectable orchestration runtime state |
| Process | Durable Workflows | Repeated agent processes become saved, resumable workflows instead of terminal choreography |
| Operational behavior | Team Capabilities | Team standards become reusable agent behavior instead of repeated prompts |
| Coordination | Multi-Agent Orchestration | Specialized agents hand work off through explicit workflow control |
| Governance | Reviewable AI Workflows | Agent output gets evidence-backed review, scoring, and approval gates |
Start with workflows
See Durable Workflows for the product value, then run your first workflow without writing YAML.