nodebench-mcp 2.67.0 → 2.68.0
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/dist/benchmarks/chainEval.d.ts +21 -0
- package/dist/benchmarks/chainEval.js +683 -0
- package/dist/benchmarks/chainEval.js.map +1 -0
- package/dist/benchmarks/llmJudgeEval.js +90 -7
- package/dist/benchmarks/llmJudgeEval.js.map +1 -1
- package/dist/benchmarks/pipelineEval.d.ts +63 -0
- package/dist/benchmarks/pipelineEval.js +1053 -0
- package/dist/benchmarks/pipelineEval.js.map +1 -0
- package/dist/benchmarks/searchQualityEval.js +4 -4
- package/dist/benchmarks/searchQualityEval.js.map +1 -1
- package/package.json +1 -1
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* chainEval.ts — Multi-step chain eval for NodeBench.
|
|
3
|
+
*
|
|
4
|
+
* Tests REAL agent workflows where Tool A's output feeds into Tool B's input.
|
|
5
|
+
* Single-tool eval proves "can each tool respond."
|
|
6
|
+
* Chain eval proves "can the pipeline produce a real result."
|
|
7
|
+
*
|
|
8
|
+
* 8 canonical chains matching the dogfood runbook:
|
|
9
|
+
* 1. Founder Weekly Reset (5 steps)
|
|
10
|
+
* 2. Pre-Delegation Packet (4 steps)
|
|
11
|
+
* 3. Important-Change Review (4 steps)
|
|
12
|
+
* 4. Competitor Intelligence (5 steps)
|
|
13
|
+
* 5. Banker Company Search (4 steps)
|
|
14
|
+
* 6. Student Strategy Brief (4 steps)
|
|
15
|
+
* 7. Setup + Discovery (3 steps)
|
|
16
|
+
* 8. Operator Causal Replay (4 steps)
|
|
17
|
+
*
|
|
18
|
+
* Usage:
|
|
19
|
+
* GEMINI_API_KEY=... npx tsx src/benchmarks/chainEval.ts [--chains N]
|
|
20
|
+
*/
|
|
21
|
+
export {};
|