[souravC01]

Summary

Resolves an issue where selecting a skill in the application (from Skills Matrix, ATS Keyword Bank, or Resume Best-Fit Ranker) failed to target the selected skill when generating resume bullet points, often generating bullets for the previously opened skill or defaulting to the first skill in the job posting. In addition, this change ensures candidate resume evidence is reliably supplied for grounding and rebrands the feature from "BulletSkill 2.0" to BulletCraft.

---

Root Cause & Changes

1. State Reconciliation Race in Modal (BulletOptimizerModal.jsx):

   • Root Cause: Conflicting dual useEffect hooks caused handleGenerate to execute against closed-over stale state from previous runs before asynchronous state updates took effect.
   • Fix: Replaced dual effects with lazy state initializers and a single mount trigger. Parametrized handleGenerate(e, overrideBullet, targetKeywords, targetSectionType) to accept explicit parameters rather than relying on state closures.
   • Section Toggling: Toggling between "Project Bullet" and "Work History" now immediately triggers generation with the new section type.

2. Component Lifecycle Isolation (App.jsx):

   • Conditionally mounted {isOptimizerOpen && <BulletOptimizerModal ... />} with a dynamic key incorporating active keywords and section type (optimizer-${userId}-${keywords}-${sectionType}).
   • Added a safe fallback in handleOpenOptimizer to ensure candidate evidence (resumes[0]) is passed if selectedResumeForJob was not previously set.

3. Job Analyzer Action UX (JobAnalyzer.jsx):

   • Rebranded the header action button to BulletCraft.
   • Updated the button click handler to target all required role qualifications rather than hardcoding required_skills[0].

4. Branding Updates:

   • Updated modal title to BulletCraft Resume Optimizer.
   • Updated offline mode description copy in SettingsModal.jsx to refer to "BulletCraft templates".

---

Testing & Validation

• Vitest (Frontend):
  • Added frontend/src/test/BulletOptimizerFlow.test.jsx covering immediate mount generation, candidate evidence forwarding, section type toggling, and sequential skill selections.
  • Ran npm test -- --run: 13 passed, 36 passed total.
• Frontend Build:
  • Ran npm run build: built in 5.61s without errors or warnings.
• Pytest (Backend):
  • Ran python -m pytest tests/: 205 passed in 31.85s.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
jobhelperguru	Ready	Preview	Sep 12, 2026 5:16pm UTC