# Resume Tailoring - Workflow Phases

## Phase 0: Library Initialization

**Always runs first - builds fresh resume database**

1. **Locate resume directory** (user provides path or default `./resumes/`)
2. **Scan for markdown files** using Glob tool
3. **Parse each resume:** Extract roles, bullets, skills, education
4. **Build experience database:**

```json
{
  "roles": [
    {
      "role_id": "company_title_year",
      "company": "Company Name",
      "title": "Job Title",
      "dates": "YYYY-YYYY",
      "bullets": [
        {
          "text": "Full bullet text",
          "themes": ["leadership", "technical"],
          "metrics": ["17x improvement", "$3M revenue"],
          "source_resumes": ["resume1.md"]
        }
      ]
    }
  ],
  "skills": { "technical": [], "product": [], "leadership": [] },
  "education": [],
  "user_preferences": {
    "typical_length": "1-page|2-page",
    "section_order": ["summary", "experience", "education"],
    "bullet_style": "pattern"
  }
}
```

5. **Auto-tag content:** themes, metrics, keywords

**Output:** In-memory database ready for matching

---

## Phase 1: Research Phase

**Goal:** Build comprehensive "success profile" beyond just the job description

**1.1 Job Description Parsing:**
Extract requirements, keywords, implicit preferences, red flags, role archetype (see `research-prompts.md`)

**1.2 Company Research:**
WebSearch for mission/values/culture, engineering blog, recent news

**1.3 Role Benchmarking:**
WebSearch LinkedIn profiles for common backgrounds, skills, terminology

**1.4 Success Profile Synthesis:**
Combine into structured profile: core requirements, valued capabilities, cultural fit signals, narrative themes, terminology map, risk factors + mitigations

**Checkpoint:** Present success profile to user for validation before proceeding.

**Output:** Validated success profile document

---

## Phase 2: Template Generation

**Goal:** Create resume structure optimized for this specific role

**2.1 Analyze resume library** for role archetypes, experience clusters, career narrative

**2.2 Role Consolidation Decision:**
- **Consolidate when:** Same company, similar responsibilities, page space constrained
- **Keep separate when:** Different companies (ALWAYS), dramatically different responsibilities

**2.3 Title Reframing Principles:**

Stay truthful to what you did, emphasize aspect most relevant to target:
1. **Emphasize different aspects:** "Graduate Researcher" → "Research Software Engineer" (if coding-heavy)
2. **Use industry-standard terminology:** "Scientist III" → "Senior Research Scientist"
3. **Add specialization when truthful:** "Engineer" → "ML Engineer" (if ML work substantial)

**Constraints:** Never claim work not done. Never inflate seniority beyond defensible. Company name and dates MUST be exact.

**2.4 Generate Template Structure:**

```markdown
## Professional Summary
[GUIDANCE: {X} sentences emphasizing {themes from success profile}]

## Key Skills
[STRUCTURE: {2-4 categories based on JD structure}]

## Professional Experience
### [ROLE 1 - Most Recent/Relevant]
[TITLE OPTIONS: A/B with rationale]
[BULLET ALLOCATION: {N} bullets based on relevance + recency]
Bullet 1: [SEEKING: {requirement type}]
...

## Education
[PLACEMENT: top if required/recent, bottom if experience-heavy]
```

**Checkpoint:** Present template with consolidation decisions, title options, and bullet allocation for user approval.

**Output:** Approved template skeleton

---

## Phase 2.5: Experience Discovery (OPTIONAL)

**Goal:** Surface undocumented experiences through conversational discovery

**Trigger after template approval if gaps identified:**
```
"I've identified {N} gaps or areas where we have weak matches.
Would you like a structured brainstorming session? (10-15 minutes)"
```

**Branching Interview Process** (see `branching-questions.md`):

1. **Open probe** per gap: "Have you worked with {skill}?" / "Tell me about times you've {demonstrated_skill}"
2. **Branch on answer:** YES → deep dive (scale, challenges, metrics) | INDIRECT → explore transferability | ADJACENT → explore related | NO → broader category or move on
3. **Follow-up systematically:** what, how, why → quantify → contextualize → validate
4. **Capture immediately** as structured experience with gap mapping

**Integration Options per discovery:**
1. ADD TO CURRENT RESUME
2. ADD TO LIBRARY ONLY
3. REFINE FURTHER
4. DISCARD

**Important:** Keep truthfulness bar high. Time-box to 10-15 minutes. User can skip entirely.

**Output:** New experiences integrated into library

---

## Phase 3: Assembly Phase

**Goal:** Fill approved template with best-matching content, with transparent scoring

**3.1 For Each Template Slot:**

1. Extract all candidate bullets from library + discovered experiences
2. Score each candidate (see `matching-strategies.md`):
   - Direct match (40%): Keywords, domain, technology, outcome
   - Transferable (30%): Same capability, different context
   - Adjacent (20%): Related tools, methods, problem space
   - Impact (10%): Achievement type alignment
3. Rank by score, group by confidence band: DIRECT (90-100%), TRANSFERABLE (75-89%), ADJACENT (60-74%), WEAK (<60%)
4. Present top 3 matches with analysis and recommendation

**3.2 Handle Gaps (confidence <60%):**
Options: reframe best available, acknowledge in cover letter, omit bullet slot, use best available with disclosure

**3.3 Content Reframing** (when >60% match but terminology misaligned):
Show before/after with truthfulness justification

**Checkpoint:** Present complete mapping with coverage summary, reframings applied, gaps identified. Wait for user approval.

**Output:** Complete bullet-by-bullet mapping with confidence scores

---

## Phase 4: Generation Phase

**Goal:** Create professional multi-format outputs

**4.1 Markdown Generation:**
Compile mapped content using user's formatting preferences (style, bullet structure, section order, length).

**Output:** `{Name}_{Company}_{Role}_Resume.md`

**4.2 DOCX Generation:**
Use `document-skills:docx` sub-skill. Professional fonts (Calibri 11pt), proper spacing, clean bullet formatting, header with contact info.

**Output:** `{Name}_{Company}_{Role}_Resume.docx`

**4.3 PDF Generation (Optional):**
Convert DOCX to PDF if requested.

**Output:** `{Name}_{Company}_{Role}_Resume.pdf`

**4.4 Generation Summary Report:**
Metadata file with target role summary, success profile, content mapping summary, reframings applied, source resumes used, gaps addressed, interview prep recommendations.

**Output:** `{Name}_{Company}_{Role}_Resume_Report.md`

**Present all files to user with quality metrics (JD coverage %, direct matches %, newly discovered experiences).**

---

## Phase 5: Library Update (CONDITIONAL)

**After user reviews generated resume:**

**Option 1 - Save to library:** Move files to library directory, rebuild database, preserve generation metadata.

**Option 2 - Need revisions:** Collect feedback, make changes, re-present.

**Option 3 - Save but don't add to library:** Keep files in current directory only.

**Benefits of library update:** Grows library with each resume, new bullet variations available, reframings reusable, discovered experiences permanently captured.

**Output:** Updated library database + metadata preservation (if Option 1)