resume-feedback

name	resume-feedback
description	Comprehensive developer resume feedback orchestrator. Scans the 'apps/portfolio/content' directory dynamically, reads resume Markdown/YAML files, aligns with 2026 developer recruitment trends, and generates a structured feedback report. Trigger this skill whenever the user asks for '이력서 피드백', 'resume feedback', '이력서 검토', '이력서 첨삭', '포트폴리오 피드백'.

Developer Resume Feedback Orchestrator

This skill orchestrates a multi-agent team (resume_analyzer and soft_skill_tuner) to analyze and tune developer resumes located in apps/portfolio/content. It dynamically adapts to the directory structure and cross-references 2026 recruitment trends to deliver a high-impact, actionable feedback report.

---

Workflow Phases

Phase 1: Dynamic Content Discovery & Trend Loading

1. Dynamic Scan: Do NOT assume a fixed file schema. Perform a recursive scan of apps/portfolio/content to discover all files.
2. Content Extraction: Read the contents of all discovered files. Map them contextually based on their contents:
   • Identify introductory bios / about-me statements.
   • Identify skills list.
   • Identify career history (companies, roles, dates).
   • Identify projects (descriptions, tech stacks, outcomes).
   • Identify open-source or external links (GitHub, blogs).
3. Load 2026 Trends: Read the rules file rules/recruitment_trends.md to establish the evaluation baseline for 2026 IT developer hiring criteria (on-demand hiring focus, technical reasoning, quantifiable metrics, and tool-based collaboration).

Phase 2: Multi-Agent Parallel Analysis

Invoke the subagents in parallel to perform specialized reviews:

1. Invoke resume_analyzer:

   • Task: Evaluate tech stack grouping, STAR structure of projects, technical decision logic (the "Why"), optimization metrics, and developer links.
   • Input: Extracted resume content + rules/recruitment_trends.md criteria.
   • Expected Output: Technical audit pointing out gaps, tech stack inconsistencies, and missing STAR elements with Before/After code snippets.

2. Invoke soft_skill_tuner:

   • Task: Evaluate how soft skills are expressed. Convert abstract clichés (e.g., "good communicator") into concrete tool-based collaborative actions (Figma, Storybook, Jira, Scrum sprint practices, rational decision-making).
   • Input: Extracted resume content + soft skill guidelines.
   • Expected Output: Tuned sentences and rewriting recommendations for about-me and experience descriptions showing 7 core traits through actions.

Phase 3: Feedback Consolidation & Report Generation

1. Collect Outputs: Read the feedback results from both subagents (from returned values or files in _workspace/).
2. Synthesize Report: Draft a unified, polished feedback report in Korean, organized as follows:
   • Executive Summary (핵심 역량): 3~5 lines summarizing career achievements, tech stack expertise, and business metrics.
   • 2026 Recruitment Trend Alignment: How the current resume fits the 2026 hiring trends (on-demand focus, AI screening compatibility).
   • Technical Audit & STAR Project Review: Concrete feedback on projects, showing gaps in S/T/A/R and suggesting precise performance metrics (quantification) and technical decision reasoning.
   • Soft Skills & Collaboration Tuning: Before/After comparisons converting abstract soft skills into tool-based engineering actions.
   • Missing Essentials Check: Status of GitHub, Blog, and Open Source links.
3. Output Artifact: Write the final report to _workspace/final_resume_feedback.md and display it to the user.

---

Error Handling

• If a subagent fails, do NOT crash the pipeline. Try a single retry. If it fails again, proceed using the orchestrator's own logic to fill the gaps, listing the limitation in the final report.
• If no resume files are found in apps/portfolio/content, prompt the user to specify the location of their resume or provide a sample template.

---

Test Scenarios

Standard Success Flow

1. User requests: /resume-feedback or "이력서 피드백해줘".
2. Orchestrator scans apps/portfolio/content -> Finds files.
3. Spawns resume_analyzer and soft_skill_tuner in parallel.
4. Aggregates results -> Outputs final_resume_feedback.md.

Missing Files Flow

1. User requests: "이력서 첨삭해줘" but the target folder is empty.
2. Orchestrator alerts the user, explains the expected directory layout, and asks for a file path or direct input.