@wix/evalforge-evaluator 0.75.0 → 0.77.0
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/build/index.js +6 -2
- package/build/index.js.map +2 -2
- package/build/index.mjs +6 -2
- package/build/index.mjs.map +2 -2
- package/package.json +2 -2
package/build/index.mjs
CHANGED
|
@@ -280,9 +280,12 @@ async function fetchEvaluationData(api, projectId2, evalRunId2) {
|
|
|
280
280
|
if (evalRun.skillsGroupId) {
|
|
281
281
|
skillsGroup = await api.getSkillsGroup(projectId2, evalRun.skillsGroupId);
|
|
282
282
|
if (skillsGroup.skillIds.length > 0) {
|
|
283
|
-
|
|
283
|
+
const fetchResults = await Promise.allSettled(
|
|
284
284
|
skillsGroup.skillIds.map((id) => api.getSkill(projectId2, id))
|
|
285
285
|
);
|
|
286
|
+
skills = fetchResults.filter(
|
|
287
|
+
(r) => r.status === "fulfilled"
|
|
288
|
+
).map((r) => r.value).filter((s) => !s.deleted);
|
|
286
289
|
}
|
|
287
290
|
if (evalRun.skillVersions && Object.keys(evalRun.skillVersions).length > 0) {
|
|
288
291
|
skills = await Promise.all(
|
|
@@ -1127,7 +1130,8 @@ IMPORTANT: This is an automated evaluation run. Follow these guidelines:
|
|
|
1127
1130
|
3. Do NOT use the Task tool to delegate simple operations - do them directly yourself.
|
|
1128
1131
|
4. Keep your approach simple and direct - avoid excessive planning.
|
|
1129
1132
|
5. Make targeted edits using Read and Edit tools rather than exploring the entire codebase.
|
|
1130
|
-
6. If you encounter an error, fix it directly rather than starting over
|
|
1133
|
+
6. If you encounter an error, fix it directly rather than starting over.
|
|
1134
|
+
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.`;
|
|
1131
1135
|
const fullPrompt = scenario.triggerPrompt + evaluatorPromptSuffix;
|
|
1132
1136
|
for await (const message of query({
|
|
1133
1137
|
prompt: fullPrompt,
|