agentv 2.13.0 → 2.14.1

package/README.md

@@ -60,7 +60,7 @@ tests:
 
     assert:
       - name: math_check
-        type: code_judge
+        type: code-judge
         script: ./validators/check_math.py
 ```
 
@@ -154,10 +154,10 @@ Optional sidecar YAML metadata file (`dataset.eval.yaml` alongside `dataset.json
 description: Math evaluation dataset
 dataset: math-tests
 execution:
-  target: azure_base
+  target: azure-base
 assert:
   - name: correctness
-    type: llm_judge
+    type: llm-judge
     prompt: ./judges/correctness.md
 ```
 
@@ -175,7 +175,7 @@ agentv validate evals/my-eval.yaml
 agentv eval evals/my-eval.yaml
 
 # Override target
-agentv eval --target azure_base evals/**/*.yaml
+agentv eval --target azure-base evals/**/*.yaml
 
 # Run specific test
 agentv eval --test-id case-123 evals/my-eval.yaml
@@ -219,7 +219,7 @@ Reference evaluators in your eval file:
 ```yaml
 assert:
   - name: my_validator
-    type: code_judge
+    type: code-judge
     script: ./validators/check_answer.py
 ```
 
@@ -339,7 +339,7 @@ Define execution targets in `.agentv/targets.yaml` to decouple evals from provid
 
 ```yaml
 targets:
-  - name: azure_base
+  - name: azure-base
     provider: azure
     endpoint: ${{ AZURE_OPENAI_ENDPOINT }}
     api_key: ${{ AZURE_OPENAI_API_KEY }}
@@ -348,12 +348,12 @@ targets:
   - name: vscode_dev
     provider: vscode
     workspace_template: ${{ WORKSPACE_PATH }}
-    judge_target: azure_base
+    judge_target: azure-base
 
   - name: local_agent
     provider: cli
     command: 'python agent.py --prompt-file {PROMPT_FILE} --output {OUTPUT_FILE}'
-    judge_target: azure_base
+    judge_target: azure-base
 ```
 
 Supports: `azure`, `anthropic`, `gemini`, `codex`, `copilot`, `pi-coding-agent`, `claude`, `vscode`, `vscode-insiders`, `cli`, and `mock`.
@@ -398,12 +398,12 @@ All assertions support `weight`, `required`, and `negate` flags. Use `negate: tr
 ```yaml
 assert:
   # Case-insensitive matching for natural language variation
-  - type: icontains_any
+  - type: icontains-any
     value: ["missing rule code", "need rule code", "provide rule code"]
     required: true
 
   # Multiple required terms
-  - type: icontains_all
+  - type: icontains-all
     value: ["country code", "rule codes"]
 
   # Case-insensitive regex
@@ -423,10 +423,10 @@ targets:
   # Agent target — requires judge_target for LLM-based evaluation
   - name: codex_local
     provider: codex
-    judge_target: azure_base  # Required: LLM provider for judging
+    judge_target: azure-base  # Required: LLM provider for judging
 
   # LLM target — no judge_target needed (judges itself)
-  - name: azure_base
+  - name: azure-base
     provider: azure
 ```
 
@@ -445,7 +445,7 @@ Create markdown judge files with evaluation criteria and scoring guidelines:
 ```yaml
 assert:
   - name: semantic_check
-    type: llm_judge
+    type: llm-judge
     prompt: ./judges/correctness.md
 ```
 
@@ -487,7 +487,7 @@ Configure automatic retry with exponential backoff:
 
 ```yaml
 targets:
-  - name: azure_base
+  - name: azure-base
     provider: azure
     max_retries: 5
     retry_initial_delay_ms: 2000

package/dist/{chunk-UWDI4UVN.js → chunk-5646K2XJ.js}

@@ -25,12 +25,12 @@ import {
   subscribeToCopilotCliLogEntries,
   subscribeToCopilotSdkLogEntries,
   subscribeToPiLogEntries
-} from "./chunk-FSBZM3HT.js";
+} from "./chunk-OQN2GDEU.js";
 
 // package.json
 var package_default = {
   name: "agentv",
-  version: "2.13.0",
+  version: "2.14.1",
   description: "CLI entry point for AgentV",
   type: "module",
   repository: {
@@ -1266,16 +1266,16 @@ function inferFileTypeFromPath(filePath) {
 var ASSERTION_TYPES_WITH_STRING_VALUE = /* @__PURE__ */ new Set([
   "contains",
   "icontains",
-  "starts_with",
-  "ends_with",
+  "starts-with",
+  "ends-with",
   "equals",
   "regex"
 ]);
 var ASSERTION_TYPES_WITH_ARRAY_VALUE = /* @__PURE__ */ new Set([
-  "contains_any",
-  "contains_all",
-  "icontains_any",
-  "icontains_all"
+  "contains-any",
+  "contains-all",
+  "icontains-any",
+  "icontains-all"
 ]);
 var VALID_TEST_FILE_EXTENSIONS = /* @__PURE__ */ new Set([".yaml", ".yml", ".jsonl"]);
 var NAME_PATTERN = /^[a-z0-9-]+$/;
@@ -1641,8 +1641,8 @@ function validateAssertArray(assertField, parentLocation, filePath, errors) {
       });
       continue;
     }
-    const typeValue = item.type;
-    if (typeValue === void 0 || typeof typeValue !== "string") {
+    const rawTypeValue = item.type;
+    if (rawTypeValue === void 0 || typeof rawTypeValue !== "string") {
       errors.push({
         severity: "warning",
         filePath,
@@ -1651,12 +1651,13 @@ function validateAssertArray(assertField, parentLocation, filePath, errors) {
       });
       continue;
     }
+    const typeValue = rawTypeValue.replace(/_/g, "-");
     if (!isEvaluatorKind(typeValue)) {
       errors.push({
         severity: "warning",
         filePath,
         location: `${location}.type`,
-        message: `Unknown assertion type '${typeValue}'.`
+        message: `Unknown assertion type '${rawTypeValue}'.`
       });
       continue;
     }
@@ -1881,7 +1882,7 @@ var MOCK_SETTINGS = /* @__PURE__ */ new Set([
   "delayMinMs",
   "delayMaxMs",
   "trace"
-  // For testing tool_trajectory evaluator
+  // For testing tool-trajectory evaluator
 ]);
 var CLAUDE_SETTINGS = /* @__PURE__ */ new Set([
   ...COMMON_SETTINGS,
@@ -3011,7 +3012,7 @@ async function runEvalCommand(input) {
   const useFileExport = !!(options.otelFile || options.traceFile);
   if (options.exportOtel || useFileExport) {
     try {
-      const { OtelTraceExporter, OTEL_BACKEND_PRESETS } = await import("./dist-CCUHG3SN.js");
+      const { OtelTraceExporter, OTEL_BACKEND_PRESETS } = await import("./dist-QR5OZ4DH.js");
       let endpoint = process.env.OTEL_EXPORTER_OTLP_ENDPOINT;
       let headers = {};
       if (options.otelBackend) {
@@ -3293,4 +3294,4 @@ export {
   selectTarget,
   runEvalCommand
 };
-//# sourceMappingURL=chunk-UWDI4UVN.js.map
+//# sourceMappingURL=chunk-5646K2XJ.js.map