agentv 2.1.0 → 2.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +21 -1
- package/dist/{chunk-5BLNVACB.js → chunk-5HTT24MQ.js} +538 -308
- package/dist/chunk-5HTT24MQ.js.map +1 -0
- package/dist/cli.js +1 -1
- package/dist/index.js +1 -1
- package/dist/templates/.claude/skills/agentv-eval-builder/SKILL.md +23 -1
- package/package.json +1 -1
- package/dist/chunk-5BLNVACB.js.map +0 -1
package/dist/cli.js
CHANGED
package/dist/index.js
CHANGED
|
@@ -7,7 +7,7 @@ description: Create and maintain AgentV YAML evaluation files for testing AI age
|
|
|
7
7
|
|
|
8
8
|
## Schema Reference
|
|
9
9
|
- Schema: `references/eval-schema.json` (JSON Schema for validation and tooling)
|
|
10
|
-
- Format: YAML
|
|
10
|
+
- Format: YAML or JSONL (see below)
|
|
11
11
|
- Examples: `references/example-evals.md`
|
|
12
12
|
|
|
13
13
|
## Feature Reference
|
|
@@ -30,6 +30,28 @@ description: Create and maintain AgentV YAML evaluation files for testing AI age
|
|
|
30
30
|
- Attachments (type: `file`) should default to the `user` role
|
|
31
31
|
- File paths: Relative (from eval file dir) or absolute with "/" prefix (from repo root)
|
|
32
32
|
|
|
33
|
+
## JSONL Format
|
|
34
|
+
|
|
35
|
+
For large-scale evaluations, use JSONL (one eval case per line) instead of YAML:
|
|
36
|
+
|
|
37
|
+
**dataset.jsonl:**
|
|
38
|
+
```jsonl
|
|
39
|
+
{"id": "test-1", "expected_outcome": "Correct answer", "input_messages": [{"role": "user", "content": "What is 2+2?"}]}
|
|
40
|
+
{"id": "test-2", "expected_outcome": "Clear explanation", "input_messages": [{"role": "user", "content": [{"type": "text", "value": "Review this"}, {"type": "file", "value": "./code.py"}]}]}
|
|
41
|
+
```
|
|
42
|
+
|
|
43
|
+
**dataset.yaml (optional sidecar for defaults):**
|
|
44
|
+
```yaml
|
|
45
|
+
description: My dataset
|
|
46
|
+
dataset: my-tests
|
|
47
|
+
execution:
|
|
48
|
+
target: azure_base
|
|
49
|
+
evaluator: llm_judge
|
|
50
|
+
```
|
|
51
|
+
|
|
52
|
+
Benefits: Git-friendly diffs, streaming-compatible, easy programmatic generation.
|
|
53
|
+
Per-case fields override sidecar defaults. See `examples/features/basic-jsonl/` for complete example.
|
|
54
|
+
|
|
33
55
|
## Custom Evaluators
|
|
34
56
|
|
|
35
57
|
Configure multiple evaluators per eval case via `execution.evaluators` array.
|