API referenceEndpoints

Set application knowledge-agent policy

Requires an application credential holding the application-admin role; the credential names the application it acts for. Generated from the production route registered in `internal/features/knowledgeagent/inbound/http_application.go`.

PUT/v1/application/agents/knowledge/policy

Requires an application credential holding the application-admin role; the credential names the application it acts for. Generated from the production route registered in internal/features/knowledgeagent/inbound/http_application.go.

Authorization

bearerAuth
AuthorizationBearer <token>

Engine-issued API key or tenant-signed JWT.

In: header

Header Parameters

If-Match*string

The strong entity tag from the most recent GET or PUT, or * for "whatever is stored now". Omitting it is 428; a tag that is not current is 412. The tag is opaque — echo it, never parse it.

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

One knowledge-agent run's budget: ten positive per-run ceilings, replaced whole. Enforcement is distributed but reads one immutable request-local value — the service refuses excess datasets and derives the deadline, the run scope meters model, tool, search and spend, the hosted tools clamp results and charge evidence, and the LLM plugin clamps output and terminates on exhaustion.

Response Body

application/json

application/json

application/json

application/json

application/json

application/json

application/json

application/json

curl -X PUT "https://example.com/v1/application/agents/knowledge/policy" \  -H "If-Match: "application-application-2"" \  -H "Content-Type: application/json" \  -d '{    "max_datasets": 10,    "max_model_calls": 4,    "max_tool_calls": 6,    "max_searches": 4,    "max_results_per_search": 6,    "max_evidence_chars": 24000,    "max_read_ids": 6,    "max_output_tokens": 1500,    "max_wall_time_seconds": 45,    "max_spend_micro_usd": 250000  }'
{  "scope": "application",  "source": "application",  "version": 0,  "policy": {    "max_datasets": 10,    "max_model_calls": 4,    "max_tool_calls": 6,    "max_searches": 4,    "max_results_per_search": 6,    "max_evidence_chars": 24000,    "max_read_ids": 6,    "max_output_tokens": 1500,    "max_wall_time_seconds": 45,    "max_spend_micro_usd": 250000  },  "effective": {    "max_datasets": 10,    "max_model_calls": 4,    "max_tool_calls": 6,    "max_searches": 4,    "max_results_per_search": 6,    "max_evidence_chars": 24000,    "max_read_ids": 6,    "max_output_tokens": 1500,    "max_wall_time_seconds": 45,    "max_spend_micro_usd": 250000  }}