List Tools
Returns the catalog of tools the agent can invoke internally during POST /agent/run / POST /agent/run/stream execution. Use this to surface agent capabilities in your UI or to validate tool availability ahead of time.
No authentication required (public). The endpoint returns a static catalog that does not depend on the workspace.
Request
No headers or query parameters required.
Response
{
"tools": [
{
"id": "search",
"displayName": "Web Search",
"category": "research",
"description": "General web search with query generation, SERP selection, scraping, and summary."
},
{
"id": "x_search",
"displayName": "X Search",
"category": "research",
"description": "Search posts on X (Twitter) with optional date range and account filters."
},
{
"id": "plan",
"displayName": "Plan",
"category": "generation",
"description": "Draft a multi-step research plan for human review before executing the full workflow."
},
{
"id": "confirm_plan",
"displayName": "Confirm Plan",
"category": "hitl",
"description": "Confirm a plan draft and start execution."
},
{
"id": "mention_agent",
"displayName": "Mention Agent",
"category": "agent_integration",
"description": "Hand off work to another agent referenced in the input via agent:// mention."
}
]
}
Response fields
| Field | Type | Description |
|---|
tools[].id | string | Internal tool identifier used by the agent |
tools[].displayName | string | Human-readable label for UI display |
tools[].category | string | Category bucket. See below |
tools[].description | string | Short description of the tool’s behavior |
tools[].trigger? | string | Trigger conditions in the chat-routing prompt (when applicable) |
tools[].inputs? | string | Primary input parameters (when applicable) |
tools[].modes? | string[] | Supported modes (e.g. matrix’s create / edit / continue) |
tools[].flow? | string[] | Ordered internal workflow steps |
tools[].sseEvents? | string[] | Related SSE event names emitted by this tool |
tools[].notes? | string | Additional notes (external endpoints, post-completion hooks, etc.) |
trigger / inputs / modes / flow / sseEvents / notes are optional — only tools that carry them will include them.
Categories
| Value | Meaning |
|---|
research | Information gathering — web search, X search, browser automation, source summary, RAG recall |
generation | Structured content generation — plans, reports, matrices |
hitl | Draft confirmation / regeneration / abort tools invoked by the LLM during human-in-the-loop flows |
agent_integration | Delegation to other agents (mention_agent), skill execution, agent memory refresh |
internal | Sub-tools not routed directly from chat-routing (e.g., SERP selection, in-browser helpers) |
This endpoint is the source of truth for the agent’s tool set. Prefer the API response over any documentation that lists tools by name.
Examples
curl "https://app.snorbe.deskrex.ai/api/v1/agent/tools"
Error responses
| HTTP status | Code | Description |
|---|
| 429 | TOO_MANY_REQUESTS | Rate limit exceeded (rare) |
