Lists all skills available in the workspace, including built-in and custom skills.
Returns every skill visible to the tenant — built-in skills shipped with Fruxon plus skills the tenant has authored or imported. Use the returned skill IDs when configuring an agent's allowed skill set; agents activate skills mid-conversation by matching the user request against the skill's description.
Authorization
Bearer JWT Authorization header using the Bearer scheme. Enter 'Bearer' [space] and then your token.
In: header
Path Parameters
The tenant identifier
Response Body
curl -X GET "https://api.fruxon.com/v1/tenants/string/skills"[
{
"id": "string",
"displayName": "string",
"description": "string",
"instructions": "string",
"tools": [
{
"integrationId": "string",
"id": "string"
}
],
"resources": [
{
"assetId": "string",
"description": "string",
"vectorized": false
}
],
"type": "SYSTEM",
"createdAt": 0,
"modifiedAt": 0
}
]{
"type": "string",
"title": "string",
"status": 0,
"detail": "string",
"instance": "string",
"property1": null,
"property2": null
}{
"type": "string",
"title": "string",
"status": 0,
"detail": "string",
"instance": "string",
"property1": null,
"property2": null
}MCP JSON-RPC endpoint (sandbox)
Mirrors the surface of `/integrations/{integration}/mcp:call` — same JSON-RPC methods, same API-key auth, same rate limiting — but `tools/call` requests are dispatched to the integration simulator instead of the real upstream. Use this from eval harnesses and CI to exercise an MCP client against deterministic, side-effect-free responses. The simulator still respects `exposedToolIds`, so changing what production clients can see also changes what sandbox clients can see.
Deletes a skill.
Hard delete — the skill is removed from the tenant and detached from any agents that referenced it. There is no soft-delete or undo; re-create via `POST` if you need it back. Returns 404 if the ID does not exist or names a built-in skill.