anpord 0.1.12 → 0.1.13

package/dist/{evals-CvOTyvCX.mjs → evals-Dcc_x6jW.mjs}

@@ -1,4 +1,5 @@
 import { n as EvalJudgment, r as EvalHarness$1, t as EvalJudge } from "./eval-judges-CwTvTq7N.mjs";
+import { t as EvalValidations } from "./eval-validations-BRdoZrTQ.mjs";
 import { Schema } from "effect";
 //#region ../schema/src/domain/eval-limits.ts
 const PROMPT_LIMIT = 16384;
@@ -25,6 +26,19 @@ const EvalVariableValue = Schema.String.pipe(Schema.maxLength(VARIABLE_VALUE_LIM
 	message: () => `A variable value must be at most ${VARIABLE_VALUE_LIMIT} characters, because it is substituted into the prompt before it is quoted.`
 }));
 //#endregion
+//#region ../schema/src/domain/eval-trigger.ts
+const EvalTrigger = Schema.Struct({
+	source: Schema.Literal("dashboard", "api", "cli", "ci", "mcp"),
+	url: Schema.optional(Schema.String.pipe(Schema.maxLength(2048), Schema.filter((value) => {
+		if (!URL.canParse(value)) return false;
+		const url = new URL(value);
+		return ["https:", "http:"].includes(url.protocol) && !url.username && !url.password;
+	}, { message: () => "Use an HTTP or HTTPS URL without credentials." })))
+}).annotations({
+	identifier: "EvalTrigger",
+	description: "What started this run. Client-reported context, not authorization."
+});
+//#endregion
 //#region ../schema/src/domain/credentials.ts
 const CredentialScope = Schema.Literal("organization", "personal");
 const CredentialStatus = Schema.Literal("active", "invalid");
@@ -109,6 +123,13 @@ Schema.Struct({
 	status: Schema.Literal("pending", "complete", "failed", "expired")
 });
 //#endregion
+//#region ../schema/src/domain/eval-source-files.ts
+const EvalSourceFile = Schema.Struct({
+	path: Schema.String.pipe(Schema.maxLength(512), Schema.pattern(/^(?!\.{1,2}(?:\/|$))[^/\\:\0]+(?:\/(?!\.{1,2}(?:\/|$))[^/\\:\0]+)*$/)),
+	content: Schema.String.pipe(Schema.maxLength(1e6))
+});
+const EvalSourceFiles = Schema.Array(EvalSourceFile).pipe(Schema.maxItems(100), Schema.filter((files) => new Set(files.map(({ path }) => path)).size === files.length), Schema.filter((files) => files.reduce((size, file) => size + file.content.length, 0) <= 1e6));
+//#endregion
 //#region ../schema/src/domain/harness-profile.ts
 const PROFILE_LIMITS = {
 	fileChars: 96e3,
@@ -150,8 +171,8 @@ const PROFILE_HARNESS_RULE = "The command harness needs a profile with a run com
 //#endregion
 //#region ../schema/src/domain/evals.ts
 const EvalHarness = EvalHarness$1;
-const EvalProvider = Schema.Literal("daytona", "e2b", "upstash", "modal", "cloudflare", "vercel");
-EvalProvider.literals;
+const EvalSandbox = Schema.Literal("daytona", "e2b", "upstash", "modal", "cloudflare", "vercel");
+EvalSandbox.literals;
 const EvalTrialStatus = Schema.Literal("queued", "running", "passed", "failed", "void");
 const EvalRunStatus = Schema.Literal("running", "finished", "failed");
 const EvalSource = Schema.Union(Schema.Struct({ kind: Schema.Literal("empty") }), Schema.Struct({
@@ -169,15 +190,22 @@ const EvalSource = Schema.Union(Schema.Struct({ kind: Schema.Literal("empty") })
 	identifier: "EvalSource"
 });
 const EvalCodeValidator = Schema.Struct({
+	capture: Schema.optional(Schema.Boolean),
+	manifest: Schema.optional(Schema.Array(Schema.Struct({
+		index: Schema.NonNegativeInt,
+		name: Schema.String.pipe(Schema.maxLength(200))
+	})).pipe(Schema.minItems(1), Schema.maxItems(20), Schema.filter((checks) => new Set(checks.map((check) => check.index)).size === checks.length, { message: () => "Validator indices must be unique" }))),
 	name: Schema.String.pipe(Schema.minLength(1), Schema.maxLength(100)),
 	source: Schema.String.pipe(Schema.minLength(1), Schema.maxLength(1e6))
 });
-const EvalValidator = Schema.Union(EvalCodeValidator, Schema.Struct({
+const EvalValidator = Schema.Union(EvalCodeValidator.pipe(Schema.extend(Schema.Struct({ sourceFiles: Schema.optional(EvalSourceFiles) }))), Schema.Struct({
 	kind: Schema.Literal("judged"),
+	capture: Schema.optional(Schema.Boolean),
 	name: Schema.String.pipe(Schema.minLength(1), Schema.maxLength(100)),
 	checks: Schema.Array(EvalCodeValidator).pipe(Schema.maxItems(20)),
-	judges: Schema.Array(EvalJudge).pipe(Schema.minItems(1), Schema.maxItems(20))
-})).annotations({
+	judges: Schema.Array(EvalJudge).pipe(Schema.minItems(1), Schema.maxItems(20)),
+	sourceFiles: Schema.optional(EvalSourceFiles)
+}).pipe(Schema.filter((validator) => validator.checks.reduce((total, check) => total + (check.manifest?.length ?? 1), 0) <= 20, { message: () => "At most 20 code validators are supported" }))).annotations({
 	description: "A bundled TypeScript validator and its exported function name.",
 	identifier: "EvalValidator"
 });
@@ -227,7 +255,7 @@ const EvalTask = Schema.Struct({
 	harnessVersion: Schema.String,
 	model: Schema.String,
 	profile: Schema.optional(Schema.NullOr(EvalTaskProfile)),
-	provider: EvalProvider
+	sandbox: EvalSandbox
 }).annotations({
 	description: "The harness, installed version, profile, model, and sandbox for a cell.",
 	identifier: "EvalTask"
@@ -237,7 +265,7 @@ const EvalTaskRequest = Schema.Struct({
 	harness: EvalHarness,
 	model: Schema.String,
 	profile: Schema.optional(HarnessProfile),
-	provider: EvalProvider
+	sandbox: Schema.optional(EvalSandbox)
 }).pipe(Schema.filter(profileFitsHarness, { message: () => PROFILE_HARNESS_RULE }));
 const EvalName = Schema.String.pipe(Schema.minLength(1), Schema.maxLength(100));
 Schema.Struct({
@@ -265,6 +293,7 @@ const EvalJournalEntry = Schema.Union(Schema.Struct({
 	exitCode: Schema.NullOr(Schema.Int),
 	finishedAtMillis: OccurredAtMillis,
 	output: Schema.String,
+	outputTruncated: Schema.optional(Schema.Boolean),
 	startedAtMillis: OccurredAtMillis
 }), Schema.Struct({
 	_tag: Schema.Literal("message"),
@@ -274,7 +303,13 @@ const EvalJournalEntry = Schema.Union(Schema.Struct({
 }), Schema.Struct({
 	_tag: Schema.Literal("toolCall"),
 	finishedAtMillis: OccurredAtMillis,
+	input: Schema.optional(Schema.String),
 	name: Schema.String,
+	output: Schema.optional(Schema.String),
+	error: Schema.optional(Schema.String),
+	outputTruncated: Schema.optional(Schema.Boolean),
+	inputTruncated: Schema.optional(Schema.Boolean),
+	errorTruncated: Schema.optional(Schema.Boolean),
 	startedAtMillis: Schema.optional(OccurredAtMillis),
 	status: Schema.NullOr(Schema.String)
 }), Schema.Struct({
@@ -319,6 +354,7 @@ const EvalCosts = Schema.Struct({
 	identifier: "EvalCosts"
 });
 const EvalTrial = Schema.Struct({
+	validations: Schema.optional(EvalValidations),
 	judgments: Schema.optional(Schema.Array(EvalJudgment)),
 	commands: Schema.Int,
 	costs: Schema.NullOr(EvalCosts),
@@ -378,6 +414,7 @@ const EvalSetup = Schema.Struct({
 	repoUrl: Schema.NullOr(Schema.String),
 	prepareName: Schema.NullOr(Schema.String),
 	validatorName: Schema.NullOr(Schema.String),
+	validatorFiles: Schema.optional(EvalSourceFiles),
 	verifyCommand: Schema.NullOr(Schema.String),
 	workspace: Schema.String
 }).annotations({
@@ -405,6 +442,7 @@ const EvalTimestamp = Schema.DateTimeUtc.annotations({
 	jsonSchema: { format: "date-time" }
 });
 const EvalRun = Schema.Struct({
+	trigger: Schema.optionalWith(Schema.NullOr(EvalTrigger), { default: () => null }),
 	cases: Schema.Array(Schema.String),
 	cells: Schema.Array(EvalCell),
 	costs: Schema.NullOr(EvalCosts),
@@ -420,6 +458,7 @@ const EvalRun = Schema.Struct({
 	identifier: "EvalRun"
 });
 const EvalRunSummary = Schema.Struct({
+	trigger: Schema.optionalWith(Schema.NullOr(EvalTrigger), { default: () => null }),
 	caseCount: Schema.Int,
 	columns: Schema.Array(EvalTask),
 	commandMax: Schema.NullOr(Schema.Int),
@@ -449,6 +488,7 @@ const EvalRunPage = Schema.Struct({
 	total: Schema.Int
 });
 const EvalCellHistoryEntry = Schema.Struct({
+	trigger: Schema.optionalWith(Schema.NullOr(EvalTrigger), { default: () => null }),
 	distribution: EvalDistribution,
 	finishedAt: Schema.NullOr(EvalTimestamp),
 	harnessVersion: Schema.String,
@@ -470,7 +510,7 @@ const PlaygroundCaseView = Schema.Struct({
 const PlaygroundColumnView = Schema.Struct({
 	harness: EvalHarness,
 	model: Schema.String,
-	provider: EvalProvider
+	sandbox: EvalSandbox
 });
 const PlaygroundConfigView = Schema.Struct({
 	cases: Schema.Array(PlaygroundCaseView),
@@ -496,7 +536,7 @@ Schema.Struct({
 	connections: CredentialSelections,
 	name: Schema.String,
 	prompt: Schema.String,
-	providers: Schema.mutable(Schema.Array(EvalProvider)).pipe(Schema.minItems(1), Schema.annotations({ message: () => "Choose at least one sandbox." })),
+	sandboxes: Schema.mutable(Schema.Array(EvalSandbox)),
 	trials: Schema.Int.pipe(Schema.between(1, 10), Schema.annotations({ message: () => "Run between 1 and 10 trials." }))
 });
 Schema.Struct({
@@ -517,7 +557,10 @@ const StartedEval = Schema.Struct({ id: Schema.String }).annotations({
 	description: "The id of an eval run accepted for background execution.",
 	identifier: "StartedEval"
 });
-const RerunCellRequest = Schema.Struct({ trials: Schema.Int.pipe(Schema.between(1, 10)) });
+const RerunCellRequest = Schema.Struct({
+	trigger: Schema.optional(EvalTrigger),
+	trials: Schema.Int.pipe(Schema.between(1, 10))
+});
 const CatalogueModel = Schema.Struct({
 	displayName: Schema.String,
 	id: Schema.String,
@@ -536,4 +579,4 @@ const ModelCatalogue = Schema.Struct({
 	identifier: "ModelCatalogue"
 });
 //#endregion
-export { EvalPrompt as C, EvalCaseName as S, HarnessProfile as _, EvalPageCursor as a, ProfilePath as b, EvalRun as c, EvalValidator as d, EvalVariables as f, EnvName as g, StartedEval as h, EvalName as i, EvalRunPage as l, RerunCellRequest as m, EvalCellHistoryEntry as n, EvalPrepare as o, ModelCatalogue as p, EvalHarness as r, EvalProvider as s, CaseCache as t, EvalSource as u, PROFILE_HARNESS_RULE as v, EvalVerify as w, profileFitsHarness as x, PROFILE_LIMITS as y };
+export { EvalTrigger as C, EvalVerify as E, EvalSourceFiles as S, EvalPrompt as T, HarnessProfile as _, EvalPageCursor as a, ProfilePath as b, EvalRunPage as c, EvalValidator as d, EvalVariables as f, EnvName as g, StartedEval as h, EvalName as i, EvalSandbox as l, RerunCellRequest as m, EvalCellHistoryEntry as n, EvalPrepare as o, ModelCatalogue as p, EvalHarness as r, EvalRun as s, CaseCache as t, EvalSource as u, PROFILE_HARNESS_RULE as v, EvalCaseName as w, profileFitsHarness as x, PROFILE_LIMITS as y };

package/dist/{evals-B04M0apj.cjs → evals-DouMsrsM.cjs}

@@ -1,4 +1,5 @@
 const require_eval_judges = require("./eval-judges-BkmiVLLq.cjs");
+const require_eval_validations = require("./eval-validations-_MlnQ_uv.cjs");
 let effect = require("effect");
 //#region ../schema/src/domain/eval-limits.ts
 const PROMPT_LIMIT = 16384;
@@ -25,6 +26,19 @@ const EvalVariableValue = effect.Schema.String.pipe(effect.Schema.maxLength(VARI
 	message: () => `A variable value must be at most ${VARIABLE_VALUE_LIMIT} characters, because it is substituted into the prompt before it is quoted.`
 }));
 //#endregion
+//#region ../schema/src/domain/eval-trigger.ts
+const EvalTrigger = effect.Schema.Struct({
+	source: effect.Schema.Literal("dashboard", "api", "cli", "ci", "mcp"),
+	url: effect.Schema.optional(effect.Schema.String.pipe(effect.Schema.maxLength(2048), effect.Schema.filter((value) => {
+		if (!URL.canParse(value)) return false;
+		const url = new URL(value);
+		return ["https:", "http:"].includes(url.protocol) && !url.username && !url.password;
+	}, { message: () => "Use an HTTP or HTTPS URL without credentials." })))
+}).annotations({
+	identifier: "EvalTrigger",
+	description: "What started this run. Client-reported context, not authorization."
+});
+//#endregion
 //#region ../schema/src/domain/credentials.ts
 const CredentialScope = effect.Schema.Literal("organization", "personal");
 const CredentialStatus = effect.Schema.Literal("active", "invalid");
@@ -109,6 +123,13 @@ effect.Schema.Struct({
 	status: effect.Schema.Literal("pending", "complete", "failed", "expired")
 });
 //#endregion
+//#region ../schema/src/domain/eval-source-files.ts
+const EvalSourceFile = effect.Schema.Struct({
+	path: effect.Schema.String.pipe(effect.Schema.maxLength(512), effect.Schema.pattern(/^(?!\.{1,2}(?:\/|$))[^/\\:\0]+(?:\/(?!\.{1,2}(?:\/|$))[^/\\:\0]+)*$/)),
+	content: effect.Schema.String.pipe(effect.Schema.maxLength(1e6))
+});
+const EvalSourceFiles = effect.Schema.Array(EvalSourceFile).pipe(effect.Schema.maxItems(100), effect.Schema.filter((files) => new Set(files.map(({ path }) => path)).size === files.length), effect.Schema.filter((files) => files.reduce((size, file) => size + file.content.length, 0) <= 1e6));
+//#endregion
 //#region ../schema/src/domain/harness-profile.ts
 const PROFILE_LIMITS = {
 	fileChars: 96e3,
@@ -150,8 +171,8 @@ const PROFILE_HARNESS_RULE = "The command harness needs a profile with a run com
 //#endregion
 //#region ../schema/src/domain/evals.ts
 const EvalHarness = require_eval_judges.EvalHarness;
-const EvalProvider = effect.Schema.Literal("daytona", "e2b", "upstash", "modal", "cloudflare", "vercel");
-EvalProvider.literals;
+const EvalSandbox = effect.Schema.Literal("daytona", "e2b", "upstash", "modal", "cloudflare", "vercel");
+EvalSandbox.literals;
 const EvalTrialStatus = effect.Schema.Literal("queued", "running", "passed", "failed", "void");
 const EvalRunStatus = effect.Schema.Literal("running", "finished", "failed");
 const EvalSource = effect.Schema.Union(effect.Schema.Struct({ kind: effect.Schema.Literal("empty") }), effect.Schema.Struct({
@@ -169,15 +190,22 @@ const EvalSource = effect.Schema.Union(effect.Schema.Struct({ kind: effect.Schem
 	identifier: "EvalSource"
 });
 const EvalCodeValidator = effect.Schema.Struct({
+	capture: effect.Schema.optional(effect.Schema.Boolean),
+	manifest: effect.Schema.optional(effect.Schema.Array(effect.Schema.Struct({
+		index: effect.Schema.NonNegativeInt,
+		name: effect.Schema.String.pipe(effect.Schema.maxLength(200))
+	})).pipe(effect.Schema.minItems(1), effect.Schema.maxItems(20), effect.Schema.filter((checks) => new Set(checks.map((check) => check.index)).size === checks.length, { message: () => "Validator indices must be unique" }))),
 	name: effect.Schema.String.pipe(effect.Schema.minLength(1), effect.Schema.maxLength(100)),
 	source: effect.Schema.String.pipe(effect.Schema.minLength(1), effect.Schema.maxLength(1e6))
 });
-const EvalValidator = effect.Schema.Union(EvalCodeValidator, effect.Schema.Struct({
+const EvalValidator = effect.Schema.Union(EvalCodeValidator.pipe(effect.Schema.extend(effect.Schema.Struct({ sourceFiles: effect.Schema.optional(EvalSourceFiles) }))), effect.Schema.Struct({
 	kind: effect.Schema.Literal("judged"),
+	capture: effect.Schema.optional(effect.Schema.Boolean),
 	name: effect.Schema.String.pipe(effect.Schema.minLength(1), effect.Schema.maxLength(100)),
 	checks: effect.Schema.Array(EvalCodeValidator).pipe(effect.Schema.maxItems(20)),
-	judges: effect.Schema.Array(require_eval_judges.EvalJudge).pipe(effect.Schema.minItems(1), effect.Schema.maxItems(20))
-})).annotations({
+	judges: effect.Schema.Array(require_eval_judges.EvalJudge).pipe(effect.Schema.minItems(1), effect.Schema.maxItems(20)),
+	sourceFiles: effect.Schema.optional(EvalSourceFiles)
+}).pipe(effect.Schema.filter((validator) => validator.checks.reduce((total, check) => total + (check.manifest?.length ?? 1), 0) <= 20, { message: () => "At most 20 code validators are supported" }))).annotations({
 	description: "A bundled TypeScript validator and its exported function name.",
 	identifier: "EvalValidator"
 });
@@ -227,7 +255,7 @@ const EvalTask = effect.Schema.Struct({
 	harnessVersion: effect.Schema.String,
 	model: effect.Schema.String,
 	profile: effect.Schema.optional(effect.Schema.NullOr(EvalTaskProfile)),
-	provider: EvalProvider
+	sandbox: EvalSandbox
 }).annotations({
 	description: "The harness, installed version, profile, model, and sandbox for a cell.",
 	identifier: "EvalTask"
@@ -237,7 +265,7 @@ const EvalTaskRequest = effect.Schema.Struct({
 	harness: EvalHarness,
 	model: effect.Schema.String,
 	profile: effect.Schema.optional(HarnessProfile),
-	provider: EvalProvider
+	sandbox: effect.Schema.optional(EvalSandbox)
 }).pipe(effect.Schema.filter(profileFitsHarness, { message: () => PROFILE_HARNESS_RULE }));
 const EvalName = effect.Schema.String.pipe(effect.Schema.minLength(1), effect.Schema.maxLength(100));
 effect.Schema.Struct({
@@ -265,6 +293,7 @@ const EvalJournalEntry = effect.Schema.Union(effect.Schema.Struct({
 	exitCode: effect.Schema.NullOr(effect.Schema.Int),
 	finishedAtMillis: OccurredAtMillis,
 	output: effect.Schema.String,
+	outputTruncated: effect.Schema.optional(effect.Schema.Boolean),
 	startedAtMillis: OccurredAtMillis
 }), effect.Schema.Struct({
 	_tag: effect.Schema.Literal("message"),
@@ -274,7 +303,13 @@ const EvalJournalEntry = effect.Schema.Union(effect.Schema.Struct({
 }), effect.Schema.Struct({
 	_tag: effect.Schema.Literal("toolCall"),
 	finishedAtMillis: OccurredAtMillis,
+	input: effect.Schema.optional(effect.Schema.String),
 	name: effect.Schema.String,
+	output: effect.Schema.optional(effect.Schema.String),
+	error: effect.Schema.optional(effect.Schema.String),
+	outputTruncated: effect.Schema.optional(effect.Schema.Boolean),
+	inputTruncated: effect.Schema.optional(effect.Schema.Boolean),
+	errorTruncated: effect.Schema.optional(effect.Schema.Boolean),
 	startedAtMillis: effect.Schema.optional(OccurredAtMillis),
 	status: effect.Schema.NullOr(effect.Schema.String)
 }), effect.Schema.Struct({
@@ -319,6 +354,7 @@ const EvalCosts = effect.Schema.Struct({
 	identifier: "EvalCosts"
 });
 const EvalTrial = effect.Schema.Struct({
+	validations: effect.Schema.optional(require_eval_validations.EvalValidations),
 	judgments: effect.Schema.optional(effect.Schema.Array(require_eval_judges.EvalJudgment)),
 	commands: effect.Schema.Int,
 	costs: effect.Schema.NullOr(EvalCosts),
@@ -378,6 +414,7 @@ const EvalSetup = effect.Schema.Struct({
 	repoUrl: effect.Schema.NullOr(effect.Schema.String),
 	prepareName: effect.Schema.NullOr(effect.Schema.String),
 	validatorName: effect.Schema.NullOr(effect.Schema.String),
+	validatorFiles: effect.Schema.optional(EvalSourceFiles),
 	verifyCommand: effect.Schema.NullOr(effect.Schema.String),
 	workspace: effect.Schema.String
 }).annotations({
@@ -405,6 +442,7 @@ const EvalTimestamp = effect.Schema.DateTimeUtc.annotations({
 	jsonSchema: { format: "date-time" }
 });
 const EvalRun = effect.Schema.Struct({
+	trigger: effect.Schema.optionalWith(effect.Schema.NullOr(EvalTrigger), { default: () => null }),
 	cases: effect.Schema.Array(effect.Schema.String),
 	cells: effect.Schema.Array(EvalCell),
 	costs: effect.Schema.NullOr(EvalCosts),
@@ -420,6 +458,7 @@ const EvalRun = effect.Schema.Struct({
 	identifier: "EvalRun"
 });
 const EvalRunSummary = effect.Schema.Struct({
+	trigger: effect.Schema.optionalWith(effect.Schema.NullOr(EvalTrigger), { default: () => null }),
 	caseCount: effect.Schema.Int,
 	columns: effect.Schema.Array(EvalTask),
 	commandMax: effect.Schema.NullOr(effect.Schema.Int),
@@ -449,6 +488,7 @@ const EvalRunPage = effect.Schema.Struct({
 	total: effect.Schema.Int
 });
 const EvalCellHistoryEntry = effect.Schema.Struct({
+	trigger: effect.Schema.optionalWith(effect.Schema.NullOr(EvalTrigger), { default: () => null }),
 	distribution: EvalDistribution,
 	finishedAt: effect.Schema.NullOr(EvalTimestamp),
 	harnessVersion: effect.Schema.String,
@@ -470,7 +510,7 @@ const PlaygroundCaseView = effect.Schema.Struct({
 const PlaygroundColumnView = effect.Schema.Struct({
 	harness: EvalHarness,
 	model: effect.Schema.String,
-	provider: EvalProvider
+	sandbox: EvalSandbox
 });
 const PlaygroundConfigView = effect.Schema.Struct({
 	cases: effect.Schema.Array(PlaygroundCaseView),
@@ -496,7 +536,7 @@ effect.Schema.Struct({
 	connections: CredentialSelections,
 	name: effect.Schema.String,
 	prompt: effect.Schema.String,
-	providers: effect.Schema.mutable(effect.Schema.Array(EvalProvider)).pipe(effect.Schema.minItems(1), effect.Schema.annotations({ message: () => "Choose at least one sandbox." })),
+	sandboxes: effect.Schema.mutable(effect.Schema.Array(EvalSandbox)),
 	trials: effect.Schema.Int.pipe(effect.Schema.between(1, 10), effect.Schema.annotations({ message: () => "Run between 1 and 10 trials." }))
 });
 effect.Schema.Struct({
@@ -517,7 +557,10 @@ const StartedEval = effect.Schema.Struct({ id: effect.Schema.String }).annotatio
 	description: "The id of an eval run accepted for background execution.",
 	identifier: "StartedEval"
 });
-const RerunCellRequest = effect.Schema.Struct({ trials: effect.Schema.Int.pipe(effect.Schema.between(1, 10)) });
+const RerunCellRequest = effect.Schema.Struct({
+	trigger: effect.Schema.optional(EvalTrigger),
+	trials: effect.Schema.Int.pipe(effect.Schema.between(1, 10))
+});
 const CatalogueModel = effect.Schema.Struct({
 	displayName: effect.Schema.String,
 	id: effect.Schema.String,
@@ -590,12 +633,6 @@ Object.defineProperty(exports, "EvalPrompt", {
 		return EvalPrompt;
 	}
 });
-Object.defineProperty(exports, "EvalProvider", {
-	enumerable: true,
-	get: function() {
-		return EvalProvider;
-	}
-});
 Object.defineProperty(exports, "EvalRun", {
 	enumerable: true,
 	get: function() {
@@ -608,12 +645,30 @@ Object.defineProperty(exports, "EvalRunPage", {
 		return EvalRunPage;
 	}
 });
+Object.defineProperty(exports, "EvalSandbox", {
+	enumerable: true,
+	get: function() {
+		return EvalSandbox;
+	}
+});
 Object.defineProperty(exports, "EvalSource", {
 	enumerable: true,
 	get: function() {
 		return EvalSource;
 	}
 });
+Object.defineProperty(exports, "EvalSourceFiles", {
+	enumerable: true,
+	get: function() {
+		return EvalSourceFiles;
+	}
+});
+Object.defineProperty(exports, "EvalTrigger", {
+	enumerable: true,
+	get: function() {
+		return EvalTrigger;
+	}
+});
 Object.defineProperty(exports, "EvalValidator", {
 	enumerable: true,
 	get: function() {

package/dist/index.cjs

@@ -1,5 +1,5 @@
 Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
-const require_client = require("./client-BnimnYiJ.cjs");
+const require_client = require("./client-C-P2I00z.cjs");
 const require_errors = require("./errors-BX1wry8K.cjs");
 const require_source = require("./source-DvlGawEw.cjs");
 let _effect_platform = require("@effect/platform");

package/dist/index.d.cts

@@ -1,6 +1,6 @@
-import { a as layer, i as DEFAULT_BASE_URL, n as AnpordClient, o as make, r as ClientOptions, t as AnpordApi } from "./client-CFTj9ad0.cjs";
-import { _ as EvalVerdict, a as EvalDistribution, b as StartedEval, c as EvalProvider, d as EvalRunSummary, f as EvalSource, g as EvalUsage, h as EvalTrialStatus, i as EvalComparison, l as EvalRun, m as EvalTrial, n as EvalCell, o as EvalHarness, p as EvalTask, r as EvalCellHistoryEntry, s as EvalJournalEntry, t as CatalogueModel, u as EvalRunStatus, v as ModelCatalogue, y as RerunCellRequest } from "./evals-BUBzpke_.cjs";
-import { a as EvalTaskDefinition, c as Prepare, d as ProfileRef, f as Validator, i as EvalDefinition, l as PrepareContext, m as ValidatorResult, n as CommandResult, o as ExecOptions, p as ValidatorContext, r as EvalCaseDefinition, s as HarnessRef, t as CaseCache, u as PrepareValue } from "./types-nr4-1c3W.cjs";
+import { a as layer, i as DEFAULT_BASE_URL, n as AnpordClient, o as make, r as ClientOptions, t as AnpordApi } from "./client-BirFO3jz.cjs";
+import { _ as EvalVerdict, a as EvalDistribution, b as StartedEval, c as EvalRun, d as EvalSandbox, f as EvalSource, g as EvalUsage, h as EvalTrialStatus, i as EvalComparison, l as EvalRunStatus, m as EvalTrial, n as EvalCell, o as EvalHarness, p as EvalTask, r as EvalCellHistoryEntry, s as EvalJournalEntry, t as CatalogueModel, u as EvalRunSummary, v as ModelCatalogue, y as RerunCellRequest } from "./evals-B6aT3xgH.cjs";
+import { a as EvalTaskDefinition, c as Prepare, d as ProfileRef, f as Validator, i as EvalDefinition, l as PrepareContext, m as ValidatorResult, n as CommandResult, o as ExecOptions, p as ValidatorContext, r as EvalCaseDefinition, s as HarnessRef, t as CaseCache, u as PrepareValue } from "./types-ISsteq95.cjs";
 import { t as McpCall } from "./calls-DYqBi5FA.cjs";
 import { t as CliCall } from "./calls-Drj_dZpu.cjs";
 import { empty, files, repo } from "./source.cjs";
@@ -128,6 +128,130 @@ declare class Anpord {
   private resolve;
 }
 //#endregion
+//#region ../schema/src/domain/eval-trigger.d.ts
+declare const EvalTrigger: Schema.Struct<{
+  source: Schema.Literal<["dashboard", "api", "cli", "ci", "mcp"]>;
+  url: Schema.optional<Schema.filter<Schema.filter<typeof Schema.String>>>;
+}>;
+type EvalTrigger = typeof EvalTrigger.Type;
+//#endregion
+//#region ../schema/src/domain/eval-validations.d.ts
+declare const ValidationValue: Schema.Struct<{
+  text: Schema.filter<typeof Schema.String>;
+  format: Schema.Literal<["text", "json"]>;
+  state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+  truncated: typeof Schema.Boolean;
+}>;
+type ValidationValue = typeof ValidationValue.Type;
+declare const ValidationCall: Schema.Struct<{
+  index: Schema.refine<number, typeof Schema.NonNegative>;
+  method: Schema.Literal<["answer", "transcript", "readText", "exists", "exec", "cli.calls", "mcp.calls"]>;
+  startedAt: typeof Schema.Number;
+  durationMs: Schema.NullOr<Schema.refine<number, typeof Schema.NonNegative>>;
+  input: Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>;
+  output: Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>;
+  error: Schema.NullOr<Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>>;
+}>;
+type ValidationCall = typeof ValidationCall.Type;
+declare const EvalValidation: Schema.Struct<{
+  id: Schema.filter<Schema.filter<typeof Schema.String>>;
+  index: Schema.refine<number, typeof Schema.NonNegative>;
+  name: Schema.filter<typeof Schema.String>;
+  kind: Schema.Literal<["code", "judge", "command"]>;
+  status: Schema.Literal<["queued", "running", "passed", "failed", "error", "skipped"]>;
+  startedAt: Schema.NullOr<typeof Schema.Number>;
+  durationMs: Schema.NullOr<Schema.refine<number, typeof Schema.NonNegative>>;
+  message: Schema.filter<typeof Schema.String>;
+  exitCode: Schema.NullOr<typeof Schema.Int>;
+  truncated: typeof Schema.Boolean;
+  judgment: Schema.optional<Schema.Struct<{
+    name: typeof Schema.String;
+    model: typeof Schema.String;
+    evaluator: typeof Schema.String;
+    score: Schema.NullOr<Schema.filter<typeof Schema.Number>>;
+    choice: Schema.NullOr<typeof Schema.String>;
+    reason: typeof Schema.String;
+    threshold: Schema.filter<typeof Schema.Number>;
+    durationMs: Schema.refine<number, typeof Schema.NonNegative>;
+    error: Schema.NullOr<typeof Schema.String>;
+  }>>;
+  input: Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>;
+  output: Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>;
+  error: Schema.NullOr<Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>>;
+  metadata: Schema.optional<Schema.Struct<{
+    text: Schema.filter<typeof Schema.String>;
+    format: Schema.Literal<["text", "json"]>;
+    state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+    truncated: typeof Schema.Boolean;
+  }>>;
+  calls: Schema.filter<Schema.Array$<Schema.Struct<{
+    index: Schema.refine<number, typeof Schema.NonNegative>;
+    method: Schema.Literal<["answer", "transcript", "readText", "exists", "exec", "cli.calls", "mcp.calls"]>;
+    startedAt: typeof Schema.Number;
+    durationMs: Schema.NullOr<Schema.refine<number, typeof Schema.NonNegative>>;
+    input: Schema.Struct<{
+      text: Schema.filter<typeof Schema.String>;
+      format: Schema.Literal<["text", "json"]>;
+      state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+      truncated: typeof Schema.Boolean;
+    }>;
+    output: Schema.Struct<{
+      text: Schema.filter<typeof Schema.String>;
+      format: Schema.Literal<["text", "json"]>;
+      state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+      truncated: typeof Schema.Boolean;
+    }>;
+    error: Schema.NullOr<Schema.Struct<{
+      text: Schema.filter<typeof Schema.String>;
+      format: Schema.Literal<["text", "json"]>;
+      state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+      truncated: typeof Schema.Boolean;
+    }>>;
+  }>>>;
+  logs: Schema.filter<Schema.Array$<Schema.Struct<{
+    index: Schema.refine<number, typeof Schema.NonNegative>;
+    at: typeof Schema.Number;
+    level: Schema.Literal<["stdout", "stderr"]>;
+    value: Schema.Struct<{
+      text: Schema.filter<typeof Schema.String>;
+      format: Schema.Literal<["text", "json"]>;
+      state: Schema.Literal<["captured", "disabled", "unavailable"]>;
+      truncated: typeof Schema.Boolean;
+    }>;
+  }>>>;
+}>;
+type EvalValidation = typeof EvalValidation.Type;
+//#endregion
 //#region ../schema/src/public/shapes.d.ts
 declare const PublicPrompt: Schema.Struct<{
   channel: Schema.NullOr<Schema.brand<Schema.filter<Schema.filter<Schema.filter<typeof Schema.String>>>, "ChannelName">>;
@@ -195,4 +319,4 @@ type StartEvalRequest = Parameters<EvalsSurface["start"]>[0];
 type EvalCase = StartEvalRequest["cases"][number];
 type EvalTaskRequest = StartEvalRequest["tasks"][number];
 //#endregion
-export { Anpord, AnpordApi, type AnpordClient, AnpordError, type AnpordOptions, type AnpordPromptVariables, type CacheOptions, type CaseCache, type CatalogueModel, type CliCall, type ClientOptions, type CommandResult, DEFAULT_BASE_URL, EvalAborted, EvalCase, type EvalCaseDefinition, type EvalCell, type EvalCellHistoryEntry, type EvalComparison, type EvalDefinition, type EvalDistribution, type EvalHarness, type EvalJournalEntry, type EvalProvider, type EvalRun, type EvalRunStatus, type EvalRunSummary, type EvalSource, type EvalTask, type EvalTaskDefinition, EvalTaskRequest, EvalTimeout, type EvalTrial, type EvalTrialStatus, type EvalUsage, type EvalVerdict, type EvalsSurface, type ExecOptions, type GetPromptOptions, type HarnessRef, type McpCall, MissingApiKey, type ModelCatalogue, type Prepare, type PrepareContext, type PrepareValue, type ProfileRef, type PromptFallback, type PromptMetadata, type PromptResult, type PromptsSurface, type PublicPrompt, type PublicPromptSummary, type PublicPromptWithVersions, type PublicVersion, type RerunCellRequest, StartEvalRequest, type StartedEval, type Validator, type ValidatorContext, type ValidatorResult, type WaitOptions, defineEval, empty, files, layer, make, repo };
+export { Anpord, AnpordApi, type AnpordClient, AnpordError, type AnpordOptions, type AnpordPromptVariables, type CacheOptions, type CaseCache, type CatalogueModel, type CliCall, type ClientOptions, type CommandResult, DEFAULT_BASE_URL, EvalAborted, EvalCase, type EvalCaseDefinition, type EvalCell, type EvalCellHistoryEntry, type EvalComparison, type EvalDefinition, type EvalDistribution, type EvalHarness, type EvalJournalEntry, type EvalRun, type EvalRunStatus, type EvalRunSummary, type EvalSandbox, type EvalSource, type EvalTask, type EvalTaskDefinition, EvalTaskRequest, EvalTimeout, type EvalTrial, type EvalTrialStatus, type EvalTrigger, type EvalUsage, type EvalValidation, type EvalVerdict, type EvalsSurface, type ExecOptions, type GetPromptOptions, type HarnessRef, type McpCall, MissingApiKey, type ModelCatalogue, type Prepare, type PrepareContext, type PrepareValue, type ProfileRef, type PromptFallback, type PromptMetadata, type PromptResult, type PromptsSurface, type PublicPrompt, type PublicPromptSummary, type PublicPromptWithVersions, type PublicVersion, type RerunCellRequest, StartEvalRequest, type StartedEval, type ValidationCall, type ValidationValue, type Validator, type ValidatorContext, type ValidatorResult, type WaitOptions, defineEval, empty, files, layer, make, repo };