> ## Documentation Index > Fetch the complete documentation index at: https://docs.signalmd.dev/llms.txt > Use this file to discover all available pages before exploring further. # List tests > The agent's tests, newest first, each with its gate state and latest result. Archived tests are left out unless asked for. ## OpenAPI ````yaml /api-reference/openapi.json get /v1/agents/{agentId}/tests openapi: 3.1.0 info: title: SignalMD API version: 1.0.0 description: >- SignalMD reads a voice agent's production calls and says what to fix, then tests whether the fix worked. This API exposes the same data the web app shows: agents, calls and their traces, signals, issues, versions and their comparisons, tests, and test runs. Authenticate with `Authorization: Bearer sig_...` (an API key from the API keys page). Read keys may call every `read` operation; write keys may also call `write` operations, within their monthly budget. Each operation's scope is in `x-signal-scope`. Errors share one envelope (`Error`), lists share one cursor scheme (`limit`, `cursor`, `hasMore`, `nextCursor`), and every response carries `Request-Id` and `RateLimit-*` headers. servers: - url: https://api.signalmd.dev description: Production - url: http://localhost:43124 description: Local development security: - bearerAuth: [] tags: - name: Account description: Who the token belongs to and what it may do. - name: Agents description: Voice agents connected to the workspace. - name: Calls description: Production calls, their traces and per-call signals. - name: Signals description: What SignalMD noticed across calls. - name: Issues description: 'Problems worth fixing: where, why, and what to change.' - name: Versions description: Agent versions and how each one did. - name: Comparisons description: Stored comparisons between two versions. - name: Tests description: Tests, test sets, personas and coverage. - name: Runs description: Test runs and the attempts inside them. paths: /v1/agents/{agentId}/tests: get: tags: - Tests summary: List tests description: >- The agent's tests, newest first, each with its gate state and latest result. Archived tests are left out unless asked for. operationId: listTests parameters: - schema: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f description: The agent's id. required: true description: The agent's id. name: agentId in: path - schema: type: integer minimum: 1 maximum: 100 default: 25 description: Items per page, 1 to 100. example: 25 required: false description: Items per page, 1 to 100. name: limit in: query - schema: type: string minLength: 1 description: '`nextCursor` from the previous page.' required: false description: '`nextCursor` from the previous page.' name: cursor in: query - schema: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f description: Only tests in this set. required: false description: Only tests in this set. name: testSetId in: query - schema: type: string enum: - hearing - understanding - flow - speaking - integrations description: Only tests in this lane. required: false description: Only tests in this lane. name: lane in: query - schema: type: string minLength: 1 description: Only tests for this intent slug. required: false description: Only tests for this intent slug. name: intent in: query - schema: type: string enum: - production - custom - probe description: Only tests from this source. required: false description: Only tests from this source. name: source in: query - schema: type: string enum: - passed - failed - flaky - error - drift - none description: >- Only tests whose latest attempt with their own caller has this result; none: never run. required: false description: >- Only tests whose latest attempt with their own caller has this result; none: never run. name: lastResult in: query - schema: type: string enum: - exclude - include - only default: exclude description: >- Archived tests are left out (exclude) unless asked for: include them, or only them. required: false description: >- Archived tests are left out (exclude) unless asked for: include them, or only them. name: archived in: query - schema: type: string enum: - draft - accepted description: Only drafts, or only accepted tests. required: false description: Only drafts, or only accepted tests. name: state in: query responses: '200': description: A page of tests. content: application/json: schema: $ref: '#/components/schemas/TestList' '400': description: 'Invalid request: a parameter or the body failed validation.' content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Missing, invalid or revoked credentials. content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: Not permitted for this principal. content: application/json: schema: $ref: '#/components/schemas/Error' '404': description: >- The agent, or a resource under it, does not exist or is not visible to this key. content: application/json: schema: $ref: '#/components/schemas/Error' '429': description: Rate limited. See Retry-After. content: application/json: schema: $ref: '#/components/schemas/Error' '500': description: Server error. The message is generic; quote the request id. content: application/json: schema: $ref: '#/components/schemas/Error' x-codeSamples: - lang: bash label: curl source: >- curl "https://api.signalmd.dev/v1/agents/6f1d2c3b-4a5e-4f60-8b7a-9c0d1e2f3a4b/tests" \ -H "Authorization: Bearer $SIGNAL_API_KEY" components: schemas: TestList: type: object properties: object: type: string enum: - list data: type: array items: $ref: '#/components/schemas/TestSummary' hasMore: type: boolean nextCursor: type: - string - 'null' description: Pass as `cursor` for the next page; null on the last page. required: - object - data - hasMore - nextCursor Error: type: object properties: error: type: object properties: type: $ref: '#/components/schemas/ErrorType' code: type: string example: agent_not_found description: The specific reason; stable within v1. message: type: string example: Agent not found. param: type: - string - 'null' description: The request field at fault, for a validation error. requestId: type: - string - 'null' example: req_8c1f0e2b4a6d9e3f7a5c1b0d quota: type: object properties: metric: type: string example: credits description: >- credits when the workspace's monthly credits are used up, or agents when the plan's agent limit is reached. activity: type: string enum: - observation - testing - chat example: testing description: >- With metric credits: what the request would have spent them on. observation is analysing calls, testing is test runs, chat is the copilot. requested: type: integer example: 18 description: >- With metric credits: the credits the request needed to start. limit: type: integer description: 'The plan''s limit: the month''s credits, or its agents.' used: type: integer description: How much of the limit is used this month. plan: type: string example: free_org description: The workspace's plan slug. upgradePath: type: string example: /billing description: Where in the web app the plan is changed. required: - metric - limit - used - plan - upgradePath description: >- Present when code is quota_exceeded: the plan limit the request would pass. proposalId: type: string format: uuid description: >- Present when code is criteria_locked: the pending criteria proposal the change was filed as. A person approves or rejects it. required: - type - code - message - param - requestId required: - error example: error: type: not_found_error code: agent_not_found message: Agent not found. param: null requestId: req_8c1f0e2b4a6d9e3f7a5c1b0d TestSummary: type: object properties: object: type: string enum: - test id: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f title: type: - string - 'null' description: What the test list calls it. Null on tests made before titles. source: type: string enum: - production - custom - probe description: >- production: a stored call replayed. custom: written as a spec. probe: written for a brief line no call exercises; never gates. lane: type: - string - 'null' enum: - hearing - understanding - flow - speaking - integrations - null description: >- The lane a defect lives in: hearing, understanding, flow, speaking or integrations. intentSlug: type: - string - 'null' description: The caller's intent the test covers. tier: type: string enum: - text - phone description: text runs a simulated chat; phone is not built yet. sourceCallId: type: - string - 'null' format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f description: The production call it was made from, or based on. sourceIssueKey: type: - string - 'null' description: The issue it reproduces, for a test made from an issue. persona: type: object properties: slug: type: - string - 'null' description: >- The persona slug of the test's own caller. Null when the caller had none. label: type: - string - 'null' required: - slug - label description: The caller the test plays when a run does not name personas. revision: type: integer description: Bumped by every accepted edit. GET .../revisions has each one. state: type: string enum: - draft - accepted description: >- draft: made by a key or the copilot and not accepted yet; it runs and reports and never gates or counts. accepted: counts. POST .../tests/accept accepts drafts. acceptedAt: type: - string - 'null' format: date-time example: '2026-10-02T14:03:07.120Z' description: When it was accepted. Null on a draft. criteriaApprovedAt: type: - string - 'null' format: date-time example: '2026-10-02T14:03:07.120Z' description: >- When a person approved its criteria. From then on a change to them needs a person's approval; anyone else's is filed as a criteria proposal. expect: type: string enum: - pass - fail description: >- What the test claims: an issue test expects the defect (fail) until its fix ships. guard: type: boolean description: >- Marked a guard test: it runs several times every release and must pass every time. gates: type: boolean description: >- Whether its results count toward a release. Only a calibrated test in the current scoring epoch gates. gateReason: type: string description: Why it gates or does not, in a few words. written: type: boolean description: >- SignalMD generated the caller (a generated test). Such a test reports and never gates. calibration: $ref: '#/components/schemas/TestCalibration' inRelease: type: boolean description: In the release set, which runs on every new version and from CI. testSets: type: array items: type: object properties: id: type: string format: uuid pattern: >- ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$ example: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f name: type: string required: - id - name description: The test sets it belongs to. tags: type: array items: type: string criteriaCount: type: integer description: Criteria lines after resolving against the current brief. fixtureCounts: type: object properties: total: type: integer description: Tools the test answers for. recorded: type: integer description: Answered from the source call's own recording. fromOtherCall: type: integer description: Answered from another stored call. ownerSet: type: integer description: Answered with a response someone wrote. none: type: integer description: 'No answer: the tool returns a fixed ''not available'' error.' required: - total - recorded - fromOtherCall - ownerSet - none description: Where each tool's mocked answer comes from. lastAttempt: type: - object - 'null' properties: result: type: - string - 'null' description: passed, failed, flaky, error or drift; null while unscored. startedAt: type: string format: date-time example: '2026-10-02T14:03:07.120Z' required: - result - startedAt description: >- The latest attempt with the test's own caller. Null when it never ran. autoImprove: $ref: '#/components/schemas/TestAutoImprove' archivedAt: type: - string - 'null' format: date-time example: '2026-10-02T14:03:07.120Z' description: Set when retired. An archived test is not run. createdAt: type: string format: date-time example: '2026-10-02T14:03:07.120Z' updatedAt: type: string format: date-time example: '2026-10-02T14:03:07.120Z' required: - object - id - title - source - lane - intentSlug - tier - sourceCallId - sourceIssueKey - persona - revision - state - acceptedAt - criteriaApprovedAt - expect - guard - gates - gateReason - written - calibration - inRelease - testSets - tags - criteriaCount - fixtureCounts - lastAttempt - autoImprove - archivedAt - createdAt - updatedAt example: object: test id: 3f1c2d4e-5a6b-4c7d-8e9f-0a1b2c3d4e5f title: Books a cleaning when the lookup is slow source: production lane: integrations intentSlug: book_appointment tier: text sourceCallId: 0b9e8d7c-6f5a-4b3c-9d2e-1f0a9b8c7d6e sourceIssueKey: tool_timeout:check_availability persona: slug: '@source' label: The call's own caller revision: 2 state: accepted acceptedAt: '2026-09-20T16:12:00.000Z' criteriaApprovedAt: '2026-09-20T16:12:00.000Z' expect: pass guard: false gates: true gateReason: 'Calibrated: reproduces the failure on version 7 and passes on 8.' written: true calibration: state: stable expect: pass version: 8 k: 3 'n': 3 results: - passed - passed - passed runId: d1e2f3a4-b5c6-4d7e-8f9a-0b1c2d3e4f5a revision: 2 epoch: simulatorModel: caller-setup-a judgeModel: review-setup-a judgePromptHash: 9f2c1a7e scorerVersion: '2026-09-18' at: '2026-10-02T11:07:41.000Z' reason: Passed every run on version 8. inRelease: true testSets: - id: c7d8e9f0-a1b2-4c3d-8e4f-5a6b7c8d9e0f name: Release tags: - booking criteriaCount: 2 fixtureCounts: total: 1 recorded: 1 fromOtherCall: 0 ownerSet: 0 none: 0 lastAttempt: result: passed startedAt: '2026-10-02T11:01:15.000Z' autoImprove: status: ready label: Ready for calls active: false updatedAt: '2026-09-28T09:44:02.000Z' archivedAt: null createdAt: '2026-09-20T16:12:00.000Z' updatedAt: '2026-09-28T09:40:31.000Z' ErrorType: type: string enum: - invalid_request_error - authentication_error - budget_error - permission_error - not_found_error - conflict_error - rate_limit_error - api_error TestCalibration: type: - object - 'null' properties: state: type: string enum: - uncalibrated - calibrating - reproduces - does_not_reproduce - stable - unstable - inconclusive - not_run description: >- Whether the test gates: only `reproduces` (an issue test that reliably fails on its source version) and `stable` (a test that reliably passes) do. expect: type: string enum: - pass - fail description: What calibration checked the test against. version: type: - integer - 'null' description: The agent version it was calibrated on. k: type: integer description: Runs that behaved as expected. 'n': type: integer description: Runs scored. results: type: array items: type: string description: Each calibration run's result, in order. runId: type: - string - 'null' description: The calibration run. revision: type: - integer - 'null' description: >- The test revision it measured. Null on calibrations stored before revisions were recorded. epoch: $ref: '#/components/schemas/ScoringEpoch' at: type: string description: When it was decided (ISO 8601). reason: type: string description: Why the state is what it is, in one line. required: - state - expect - version - k - 'n' - results - runId - revision - epoch - at - reason TestAutoImprove: type: object properties: status: type: - string - 'null' enum: - improving - ready - reproduces - bot_problem - needs_you - paused - stopped - null description: >- Where auto-improve stands on this test. improving: runs or a fix are in progress. ready: the test passes as written. reproduces: an issue test that captures the issue. bot_problem: the test is right and the agent fails it, confirmed on Retell. needs_you: a person has to decide, including a possible problem in the agent seen in a text run and not confirmed on Retell (run the test on Retell to confirm it). paused: waiting on the budget, a Retell key or credits (the workspace has none left this month). stopped: a person edited or archived the test. Null when it never ran on this test. label: type: - string - 'null' description: The same in one line, as the app shows it. active: type: boolean description: Whether it is still working on the test. updatedAt: type: - string - 'null' format: date-time example: '2026-10-02T14:03:07.120Z' required: - status - label - active - updatedAt ScoringEpoch: type: - object - 'null' properties: simulatorModel: type: - string - 'null' description: What played the caller. judgeModel: type: - string - 'null' description: How the criteria were scored. judgePromptHash: type: - string - 'null' description: >- Identifier of the review setup used, for comparing runs. Runs with different values were scored differently. scorerVersion: type: - string - 'null' description: The version of SignalMD's scoring. required: - simulatorModel - judgeModel - judgePromptHash - scorerVersion description: >- What scored a result. Results scored in different epochs are not compared. securitySchemes: bearerAuth: type: http scheme: bearer description: An API key (`sig_...`), or a session token from the web app. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.