> ## Documentation Index > Fetch the complete documentation index at: https://docs.humanos.tech/llms.txt > Use this file to discover all available pages before exploring further. # Agent trust window > Get what one agent did over a period of time. Answers questions like: how active was the agent, how often was it approved or denied and why, who did it transact with, and did it keep trying after its authority ended. Use your own organization DID in `orgDid` to check your agents. Use a partner organization's DID to check theirs; this requires an active sharing grant from that organization, otherwise the request returns 404. The window runs from `from` (included) to `to` (excluded), in UTC, and can cover at most 366 days. See [Risk metrics](/essentials/risk-metrics) for how sharing grants let you consult another organization's intelligence. ## OpenAPI ````yaml /versions/2026-09-01.yaml get /intelligence/window openapi: 3.1.0 info: title: Humanos API description: Humanos API version: '1.0' contact: {} servers: [] security: [] tags: - name: humanosApp description: '' paths: /intelligence/window: get: tags: - Intelligence summary: Agent trust window description: >- Get what one agent did over a period of time. Answers questions like: how active was the agent, how often was it approved or denied and why, who did it transact with, and did it keep trying after its authority ended. Use your own organization DID in `orgDid` to check your agents. Use a partner organization's DID to check theirs; this requires an active sharing grant from that organization, otherwise the request returns 404. The window runs from `from` (included) to `to` (excluded), in UTC, and can cover at most 366 days. See [Risk metrics](/essentials/risk-metrics) for how sharing grants let you consult another organization's intelligence. operationId: intelligence_window parameters: - name: orgDid required: true in: query description: >- DID of the organization whose data is consulted. Your own organization DID needs no grant; any other requires an active sharing grant from that organization to yours. schema: example: did:web:org-a.example type: string - name: did required: true in: query description: DID of the agent to measure. schema: example: did:key:z6MkhaXgBZDvotDkL5257faiztiGiC2QtKLGpbnnEGta2doK type: string - name: from required: true in: query description: Window start, inclusive (ISO 8601). schema: example: '2026-07-01T00:00:00Z' type: string - name: to required: true in: query description: >- Window end, exclusive (ISO 8601). Must be after `from`; at most 366 days later. schema: example: '2026-08-01T00:00:00Z' type: string - $ref: '#/components/parameters/ApiVersion' responses: '200': description: >- The full window body: the question echoed back, then usage and incidents. content: application/json: schema: $ref: '#/components/schemas/IntelligenceWindowEntity' '404': description: >- sharing_not_enabled — the organization does not exist or does not share with you. The two cases are deliberately indistinguishable. components: parameters: ApiVersion: name: API-Version in: header required: false schema: type: string pattern: ^\d{4}-\d{2}-\d{2}$ example: '2026-09-01' description: >- Pin request, response, and webhook shapes to a specific dated API version (YYYY-MM-DD). Omit to use the version pinned to your API key (set when the key is created; new keys default to the latest version). New integrations should target the latest version. schemas: IntelligenceWindowEntity: type: object properties: orgDid: type: string description: Organization whose data was consulted. example: did:web:org-a.example did: type: string description: Agent that was measured. from: type: string description: Window start, inclusive. example: '2026-07-01T00:00:00.000Z' to: type: string description: Window end, exclusive. example: '2026-08-01T00:00:00.000Z' computedAt: type: string description: When this report was computed. usage: description: What the agent did. allOf: - $ref: '#/components/schemas/UsageBlock' incidents: description: What went wrong. allOf: - $ref: '#/components/schemas/IncidentsBlock' required: - orgDid - did - from - to - computedAt - usage - incidents UsageBlock: type: object properties: mandates: description: The mandate request funnel over the window cohort. allOf: - $ref: '#/components/schemas/MandateFunnelBlock' presentations: $ref: '#/components/schemas/PresentationsBlock' verifications: $ref: '#/components/schemas/VerificationsBlock' verifiers: $ref: '#/components/schemas/VerifiersBlock' burnRate: $ref: '#/components/schemas/BurnRateBlock' required: - mandates - presentations - verifications - verifiers - burnRate IncidentsBlock: type: object properties: denials: description: >- Why the agent was denied, split into policy failures and lifecycle denials. allOf: - $ref: '#/components/schemas/DenialsBlock' revocations: type: object description: Mandates revoked in the window, per reason. example: user_initiated: 1 exhausted: 2 additionalProperties: type: number postRevocationActivity: $ref: '#/components/schemas/PostRevocationBlock' brokenChains: type: number description: Mandate chains currently failing cryptographic verification. required: - denials - revocations - postRevocationActivity - brokenChains MandateFunnelBlock: type: object properties: requested: type: number description: >- Mandates requested (issued) inside the window — the cohort the next three numbers partition. approved: type: number description: Of the cohort, consented by the end of the window. rejected: type: number description: Of the cohort, rejected by the end of the window. pending: type: number description: Of the cohort, still undecided at `to`. approvalRate: type: number description: approved / requested. Null when nothing was requested. nullable: true example: 0.902 required: - requested - approved - rejected - pending - approvalRate PresentationsBlock: type: object properties: total: type: number description: Presentations created in the window. verified: type: number description: Of them, how many went on to be verified inside the window. conversionRate: type: number description: verified / total. Null when there were no presentations. nullable: true example: 0.944 required: - total - verified - conversionRate VerificationsBlock: type: object properties: total: type: number description: Verifications in the window (approved + denied). approved: type: number description: Verifications that passed policy. denied: type: number description: Verifications denied. approvalRate: type: number description: approved / total. Null when the window has no verifications. nullable: true example: 0.944 required: - total - approved - denied - approvalRate VerifiersBlock: type: object properties: distinct: type: number description: Distinct counterparties that verified the agent in the window. firstSeen: type: number description: >- Of them, how many the agent had never transacted with before the window. firstSeenShare: type: number description: firstSeen / distinct. Null when the window has no verifiers. nullable: true example: 0.25 required: - distinct - firstSeen - firstSeenShare BurnRateBlock: type: object properties: approvals: type: number description: Approved verifications in the window. approvalsPerDay: type: number description: Approvals per day over the window. nullable: true example: 4.03 executionsRemaining: type: number description: >- Uses left across execution-bounded mandates live at `to`. Null when none are bounded. nullable: true projectedExhaustion: type: string description: >- Projected instant the allowance runs out at the observed rate. Null when unbounded or idle. nullable: true example: '2026-09-14T10:00:00.000Z' byAction: type: object description: >- The same figures re-cut per action, keyed by action URN — a blended projection can be wrong for every individual action. An action appears when it was used in the window or still has live mandates at `to`. Resolve a URN with GET /intelligence/action. additionalProperties: $ref: '#/components/schemas/ActionBurnBlock' example: urn:via:action:8dac0698-b005-4572-9801-7be144c545a5: approvals: 36 approvalsPerDay: 1.2 executionsRemaining: 36 projectedExhaustion: '2026-09-14T10:00:00.000Z' required: - approvals - approvalsPerDay - executionsRemaining - projectedExhaustion - byAction DenialsBlock: type: object properties: byRule: type: object description: Denials where a policy rule evaluated and failed, per rule. example: amount <= mandate.maxAmount: 5 additionalProperties: type: number byLifecycle: type: object description: >- Denials where policy never ran — the mandate was already inactive, exhausted or otherwise unusable — per reason. example: credential_inactive: 3 additionalProperties: type: number required: - byRule - byLifecycle PostRevocationBlock: type: object properties: attempts: type: number description: >- Presentation/verification events recorded AFTER a mandate's authority ended — the strongest misuse signal. mandatesInvolved: type: number description: How many mandates saw such attempts. required: - attempts - mandatesInvolved ActionBurnBlock: type: object properties: approvals: type: number description: Approved verifications on this action's mandates in the window. approvalsPerDay: type: number description: Approvals per day over the window, for this action. nullable: true example: 1.2 executionsRemaining: type: number description: >- Uses left across this action's execution-bounded mandates live at `to`. Null when none are bounded. nullable: true example: 36 projectedExhaustion: type: string description: >- Projected instant this action's allowance runs out at its observed rate. Null when unbounded or idle. nullable: true example: '2026-09-14T10:00:00.000Z' required: - approvals - approvalsPerDay - executionsRemaining - projectedExhaustion ````