Capability List
{"content":"The Capability List is your registry of encoded expertise. Every capability your team has defined, every agent assignment, every prompt binding—visible in one place. No more wondering \"does our agent know about X?\" The answer is in the table.\n\nCapabilities can be shared and versioned across a team, but they are always applied locally when an agent runs. Shared definitions, local execution.\n\n{{\u003c thumbcard src=\"/images/ui/capabilities/capability-list.png\" alt=\"Capability list with role filters, pack and category columns, prompt bindings, and agent counts\" caption=\"156 capabilities with role filters, pack grouping, and agent assignment counts\" \u003e}}\n\n## What this page covers\n\n- Registry statistics and what they mean\n- Filtering and searching capabilities\n- Table columns and how to read them\n- Creating new capabilities\n\n## Registry statistics\n\nThree cards at the top show your capability coverage:\n\n| Stat | What it shows |\n|------|---------------|\n| **Assignments** | Total agent-capability pairings (one agent + one capability = one assignment) |\n| **Unique Capability Keys** | Distinct skills tracked across your roster |\n| **Agents with Capabilities** | Active agents with at least one capability recorded |\n\nThese numbers tell you how much institutional knowledge you've encoded. More capabilities = more context that agents carry automatically.\n\n\u003e [!TIP]\n\u003e **Track growth over time.** As you encode more capabilities, watch these numbers grow. That's your team's expertise becoming agent infrastructure.\n\n## Role filter bar\n\nSix role buttons span the top, each showing a count:\n\n| Role | Purpose |\n|------|---------|\n| **Core** | Agent identity and foundational behavior. The system prompt that defines who the agent is. |\n| **Workflow** | Step-by-step procedures and execution guides. The largest category — how agents do their work. |\n| **Constraints** | Hard rules and guardrails. Build standards, coding conventions, security policies. |\n| **Context** | Reference material and domain knowledge. Background information agents can hot-load on demand. |\n| **Validation** | Quality checks, schema validation, and acceptance criteria. |\n| **Integration** | External system connectors and cross-tool workflows. |\n\nClick a role button to filter the table. Click again to clear.\n\n## Filtering capabilities\n\n### Search\n\nFind any capability instantly by:\n- Capability key: `typescript.testing.vitest`\n- Category: `orchestrator`\n- Pack: `valdr`\n- Prompt name fragment\n\nResults update as you type. No waiting, no page reloads.\n\n### Categories as signals\n\nCategories group related capabilities within a pack:\n\n| Category | What it covers |\n|----------|----------------|\n| `orchestrator` | Agent lifecycle, dispatch, and coordination |\n| `architecture` | Design patterns, system diagrams, structural decisions |\n| `registry` | Agent and capability management workflows |\n| `testing` | Test patterns, coverage requirements |\n| `operations` | Infrastructure, CI/CD, deployment |\n\nUse categories to understand your capability coverage. Missing a category? That's a gap in your encoded knowledge.\n\n## Table columns\n\n| Column | What it shows | Why it matters |\n|--------|---------------|----------------|\n| **Capability** | Display name and key identifier | Click to open the detail view |\n| **Role** | Capability role (core, workflow, constraints, context, validation, integration) | Determines precedence in prompt assembly |\n| **Pack** | Which pack this capability belongs to (`valdr`, `valdr-internal`, or custom) | Organizes capabilities by distribution source |\n| **Category** | Grouping label within the pack | Filter by domain |\n| **Prompt** | Name of the bound prompt | Links to the actual instructions the agent receives |\n| **Assigned Agents** | Count of agents using this capability | Reusability indicator |\n\n### Reading assignment counts\n\n- **High count (5+)** — Foundational capability used across many agents\n- **Medium count (2-4)** — Domain-specific capability shared by related agents\n- **Low count (1)** — Specialized capability for a specific agent role\n\nA capability with zero assignments exists but isn't being used. Consider assigning it or removing it.\n\n## Creating new capabilities\n\nClick **Add Capability** to register a new skill:\n\n1. **Key** — Lowercase identifier (e.g., `acme.ui.components`)\n2. **Category** — Optional grouping label (e.g., `frontend`)\n3. **Prompt Binding** — Select a prompt from your library\n\nThe prompt binding is critical. Without it, the capability is just a label. With it, agents receive actual instructions.\n\n\u003e [!NOTE]\n\u003e **Prompts first, capabilities second.** Create your capability prompt in the Prompts registry, then bind it to a capability. This ensures every capability has substance, not just a name.\n\n## Workflows that work\n\n### Finding capabilities by domain\n\n1. Use the search filter with a category name\n2. Browse the filtered results\n3. Check assignment counts to understand usage\n\n### Auditing capability coverage\n\n1. Sort by **Assigned Agents** ascending\n2. Look for capabilities with 0 or 1 assignments\n3. Decide: assign to more agents, or remove if unused\n\n### Discovering what an agent knows\n\n1. Note the capability key you're curious about\n2. Open the capability detail to see the Assignments tab\n3. See exactly which agents have this capability\n\n### Adding a new team convention\n\n1. Write the capability prompt in your prompts directory\n2. Register the prompt in Valdr's Prompts registry\n3. Create a capability with the key and category\n4. Bind the prompt to the capability\n5. Assign the capability to relevant agents\n\n## Pagination\n\nLarge capability libraries stay fast with pagination:\n\n- **Rows per page** — 10, 25, or 50\n- **Page navigation** — Prev/Next buttons\n- **Total count** — \"Showing 1–25 of 156\"\n\nEven with hundreds of capabilities, the list loads instantly.\n\n## Next steps\n\n- **[Capability Detail](../capability-detail/)** — See definitions, bindings, and assignments\n- **[Team Capabilities Guide](/valdr/docs/valdr-packs/team-capabilities/)** — Learn to encode your team's expertise\n\n**Open your Capability List now** and filter by category. See what knowledge you've already encoded—and identify the gaps worth filling.","description":"Your capability registry—browse, filter, and see which agents have which skills at a glance.","title":"Capability List"}
The Capability List is your registry of encoded expertise. Every capability your team has defined, every agent assignment, every prompt binding—visible in one place. No more wondering “does our agent know about X?” The answer is in the table.
Capabilities can be shared and versioned across a team, but they are always applied locally when an agent runs. Shared definitions, local execution.
156 capabilities with role filters, pack grouping, and agent assignment counts
What this page covers
- Registry statistics and what they mean
- Filtering and searching capabilities
- Table columns and how to read them
- Creating new capabilities
Registry statistics
Three cards at the top show your capability coverage:
| Stat | What it shows |
|---|---|
| Assignments | Total agent-capability pairings (one agent + one capability = one assignment) |
| Unique Capability Keys | Distinct skills tracked across your roster |
| Agents with Capabilities | Active agents with at least one capability recorded |
These numbers tell you how much institutional knowledge you’ve encoded. More capabilities = more context that agents carry automatically.
Tip
Track growth over time. As you encode more capabilities, watch these numbers grow. That’s your team’s expertise becoming agent infrastructure.
Role filter bar
Six role buttons span the top, each showing a count:
| Role | Purpose |
|---|---|
| Core | Agent identity and foundational behavior. The system prompt that defines who the agent is. |
| Workflow | Step-by-step procedures and execution guides. The largest category — how agents do their work. |
| Constraints | Hard rules and guardrails. Build standards, coding conventions, security policies. |
| Context | Reference material and domain knowledge. Background information agents can hot-load on demand. |
| Validation | Quality checks, schema validation, and acceptance criteria. |
| Integration | External system connectors and cross-tool workflows. |
Click a role button to filter the table. Click again to clear.
Filtering capabilities
Search
Find any capability instantly by:
- Capability key:
typescript.testing.vitest - Category:
orchestrator - Pack:
valdr - Prompt name fragment
Results update as you type. No waiting, no page reloads.
Categories as signals
Categories group related capabilities within a pack:
| Category | What it covers |
|---|---|
orchestrator | Agent lifecycle, dispatch, and coordination |
architecture | Design patterns, system diagrams, structural decisions |
registry | Agent and capability management workflows |
testing | Test patterns, coverage requirements |
operations | Infrastructure, CI/CD, deployment |
Use categories to understand your capability coverage. Missing a category? That’s a gap in your encoded knowledge.
Table columns
| Column | What it shows | Why it matters |
|---|---|---|
| Capability | Display name and key identifier | Click to open the detail view |
| Role | Capability role (core, workflow, constraints, context, validation, integration) | Determines precedence in prompt assembly |
| Pack | Which pack this capability belongs to (valdr, valdr-internal, or custom) | Organizes capabilities by distribution source |
| Category | Grouping label within the pack | Filter by domain |
| Prompt | Name of the bound prompt | Links to the actual instructions the agent receives |
| Assigned Agents | Count of agents using this capability | Reusability indicator |
Reading assignment counts
- High count (5+) — Foundational capability used across many agents
- Medium count (2-4) — Domain-specific capability shared by related agents
- Low count (1) — Specialized capability for a specific agent role
A capability with zero assignments exists but isn’t being used. Consider assigning it or removing it.
Creating new capabilities
Click Add Capability to register a new skill:
- Key — Lowercase identifier (e.g.,
acme.ui.components) - Category — Optional grouping label (e.g.,
frontend) - Prompt Binding — Select a prompt from your library
The prompt binding is critical. Without it, the capability is just a label. With it, agents receive actual instructions.
Note
Prompts first, capabilities second. Create your capability prompt in the Prompts registry, then bind it to a capability. This ensures every capability has substance, not just a name.
Workflows that work
Finding capabilities by domain
- Use the search filter with a category name
- Browse the filtered results
- Check assignment counts to understand usage
Auditing capability coverage
- Sort by Assigned Agents ascending
- Look for capabilities with 0 or 1 assignments
- Decide: assign to more agents, or remove if unused
Discovering what an agent knows
- Note the capability key you’re curious about
- Open the capability detail to see the Assignments tab
- See exactly which agents have this capability
Adding a new team convention
- Write the capability prompt in your prompts directory
- Register the prompt in Valdr’s Prompts registry
- Create a capability with the key and category
- Bind the prompt to the capability
- Assign the capability to relevant agents
Pagination
Large capability libraries stay fast with pagination:
- Rows per page — 10, 25, or 50
- Page navigation — Prev/Next buttons
- Total count — “Showing 1–25 of 156”
Even with hundreds of capabilities, the list loads instantly.
Next steps
- Capability Detail — See definitions, bindings, and assignments
- Team Capabilities Guide — Learn to encode your team’s expertise
Open your Capability List now and filter by category. See what knowledge you’ve already encoded—and identify the gaps worth filling.