@wix/evalforge-evaluator 0.76.0 → 0.78.0

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 CHANGED
@@ -300,9 +300,12 @@ async function fetchEvaluationData(api, projectId2, evalRunId2) {
300
300
  if (evalRun.skillsGroupId) {
301
301
  skillsGroup = await api.getSkillsGroup(projectId2, evalRun.skillsGroupId);
302
302
  if (skillsGroup.skillIds.length > 0) {
303
- skills = await Promise.all(
303
+ const fetchResults = await Promise.allSettled(
304
304
  skillsGroup.skillIds.map((id) => api.getSkill(projectId2, id))
305
305
  );
306
+ skills = fetchResults.filter(
307
+ (r) => r.status === "fulfilled"
308
+ ).map((r) => r.value).filter((s) => !s.deleted);
306
309
  }
307
310
  if (evalRun.skillVersions && Object.keys(evalRun.skillVersions).length > 0) {
308
311
  skills = await Promise.all(
@@ -1141,7 +1144,8 @@ IMPORTANT: This is an automated evaluation run. Follow these guidelines:
1141
1144
  4. Keep your approach simple and direct - avoid excessive planning.
1142
1145
  5. Make targeted edits using Read and Edit tools rather than exploring the entire codebase.
1143
1146
  6. If you encounter an error, fix it directly rather than starting over.
1144
- 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.`;
1147
+ 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.
1148
+ 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.`;
1145
1149
  const fullPrompt = scenario.triggerPrompt + evaluatorPromptSuffix;
1146
1150
  for await (const message of query({
1147
1151
  prompt: fullPrompt,