@wix/evalforge-types 0.41.0 → 0.43.0

package/build/index.mjs

@@ -949,9 +949,16 @@ var AVAILABLE_OPENAI_MODEL_IDS = Object.values(
 var OpenAIModelSchema = z4.enum(
   AVAILABLE_OPENAI_MODEL_IDS
 );
+var ALL_AVAILABLE_MODEL_IDS = [
+  ...AVAILABLE_MODEL_IDS,
+  ...AVAILABLE_OPENAI_MODEL_IDS
+];
+var AnyModelSchema = z4.enum(
+  ALL_AVAILABLE_MODEL_IDS
+);
 var nullToUndefined = (val) => val === null ? void 0 : val;
 var ModelConfigSchema = z4.object({
-  model: ClaudeModelSchema,
+  model: AnyModelSchema,
   temperature: z4.preprocess(
     nullToUndefined,
     z4.number().min(0).max(1).optional()
@@ -994,6 +1001,24 @@ var TargetSchema = TenantEntitySchema.extend({
 
 // src/target/agent.ts
 import { z as z6 } from "zod";
+var DEFAULT_EVALUATOR_SYSTEM_PROMPT = `IMPORTANT: This is an automated evaluation run. Follow these guidelines:
+1. Execute the requested changes immediately without asking for confirmation.
+2. Do NOT ask "would you like me to proceed?" or similar questions.
+3. Do NOT use the Task tool to delegate simple operations - do them directly yourself.
+4. Keep your approach simple and direct - avoid excessive planning.
+5. Make targeted edits using Read and Edit tools rather than exploring the entire codebase.
+6. If you encounter an error, fix it directly rather than starting over.
+7. Your project root is the current working directory. Always create and modify source code files relative to the project root, NOT inside .claude/skills/ directories.
+8. Before finishing, run the project's package manager install command (e.g. \`npm install\`, \`yarn install\`, or \`pnpm install\` depending on the lockfile present) to ensure all dependencies are installed and the project is ready to build.`;
+var AgentType = {
+  CLI: "cli",
+  SDK: "sdk"
+};
+var AgentTypeSchema = z6.enum([AgentType.CLI, AgentType.SDK]);
+var AGENT_TYPE_LABELS = {
+  [AgentType.CLI]: "CLI Agent",
+  [AgentType.SDK]: "Simple Agent"
+};
 var AgentRunCommand = /* @__PURE__ */ ((AgentRunCommand2) => {
   AgentRunCommand2["CLAUDE"] = "claude";
   return AgentRunCommand2;
@@ -1004,10 +1029,15 @@ var RUN_COMMAND_LABELS = {
 };
 var AgentRunCommandSchema = z6.nativeEnum(AgentRunCommand);
 var AgentSchema = TargetSchema.extend({
-  /** Command to run the agent */
-  runCommand: AgentRunCommandSchema,
+  /** Agent type: 'cli' for external CLI tools, 'sdk' for in-process SDK agents */
+  agentType: AgentTypeSchema.default(AgentType.CLI),
+  /** Command to run the agent (required for CLI agents, absent for SDK agents) */
+  runCommand: AgentRunCommandSchema.optional(),
   /** Optional model configuration override */
-  modelConfig: ModelConfigSchema.optional()
+  modelConfig: ModelConfigSchema.optional(),
+  systemPrompt: z6.string().nullish().describe(
+    "Override for eval runs. undefined=default instructions, null=raw agent, string=append to claude_code preset. See https://docs.anthropic.com/en/docs/claude-code/sdk/modifying-system-prompts"
+  )
 });
 var CreateAgentInputSchema = AgentSchema.omit({
   id: true,
@@ -1016,7 +1046,8 @@ var CreateAgentInputSchema = AgentSchema.omit({
   deleted: true
 });
 var UpdateAgentInputSchema = CreateAgentInputSchema.partial().extend({
-  modelConfig: ModelConfigSchema.optional().nullable()
+  modelConfig: ModelConfigSchema.optional().nullable(),
+  systemPrompt: z6.string().optional().nullable()
 });
 
 // src/target/skill.ts
@@ -2181,6 +2212,8 @@ function getSystemAssertion(id) {
   return SYSTEM_ASSERTIONS[id];
 }
 export {
+  AGENT_TYPE_LABELS,
+  ALL_AVAILABLE_MODEL_IDS,
   AVAILABLE_MODEL_IDS,
   AVAILABLE_OPENAI_MODEL_IDS,
   AVAILABLE_RUN_COMMANDS,
@@ -2188,7 +2221,10 @@ export {
   AgentRunCommand,
   AgentRunCommandSchema,
   AgentSchema,
+  AgentType,
+  AgentTypeSchema,
   AllowedCommands,
+  AnyModelSchema,
   ApiCallSchema,
   AssertionConfigSchema,
   AssertionParameterSchema,
@@ -2222,6 +2258,7 @@ export {
   CreateTestScenarioInputSchema,
   CreateTestSuiteInputSchema,
   CustomAssertionSchema,
+  DEFAULT_EVALUATOR_SYSTEM_PROMPT,
   DEFAULT_JUDGE_MODEL,
   DiffContentSchema,
   DiffLineSchema,