add-source

name	add-source
description	Ingests a resume, CV, or portfolio document into the ExperienceLibrary. Extracts structured experience units and merges them into CareerNavigator/ExperienceLibrary.json. Fires when the user uploads or references a new source document. Also invocable via /career-navigator:add-source.
triggers	["add this resume","add my resume","add this CV","ingest this document","add to my ExperienceLibrary","add a source","/add-source","here's my resume","upload my resume","I have a new resume","I have another resume","add this to the ExperienceLibrary"]

Extract structured experience units from a source document and merge them into the user's ExperienceLibrary.

Workflow

1. Identify the source document

If the user has provided a file path or uploaded a document — use it directly.

If no document is in context, ask:

"Which file should I add? Provide the path or paste the content directly."

Supported formats: PDF, DOCX, TXT, MD, plain text paste.

2. Read the current ExperienceLibrary

Read {user_dir}/CareerNavigator/ExperienceLibrary.json. Note:

• Existing unit IDs (to avoid collisions)
• Existing companies and roles (to detect duplicates)

If the file does not exist, create it with an empty units array before proceeding.

3. Extract experience units

Read the source document in full. For each distinct role, education entry, publication, award, or recognition, extract a structured unit:

{
  "id": "exp-{next_available_number}",
  "type": "role",
  "company": "...",
  "title": "...",
  "dates": "...",
  "summary": "1–2 sentence description of scope and context",
  "achievements": [
    {
      "id": "exp-{n}-a1",
      "theme": "impact | leadership | technical | commercial | operational",
      "text": "exact achievement text, preserved from source"
    }
  ],
  "skills": ["skill1", "skill2"],
  "performance_weight": 0.5,
  "source_document": "{filename of source}"
}

Extraction rules:

• Preserve achievement text as written — do not rephrase or improve at ingest time. Tailoring happens at resume assembly.
• Assign a theme to each achievement based on what it primarily demonstrates.
• Extract skills as they appear — normalize capitalization but do not rename or merge technologies.
• Set performance_weight to 0.5 (neutral) for all new units. The analyst agent adjusts weights from outcomes.

4. Check for duplicates

Before appending, check each extracted unit against existing ExperienceLibrary entries:

• Same company + overlapping date range = likely duplicate
• If a likely duplicate is found, surface it:

"I see an existing entry for {Role} at {Company} ({dates}) in your ExperienceLibrary. Should I replace it, merge the achievements, or add this as a separate entry?"

5. Merge into ExperienceLibrary

Append the new units to the units array in CareerNavigator/ExperienceLibrary.json. Update the meta.updated field to today's date.

Also add the source document to artifacts-index.json if it is not already listed:

{
  "id": "{uuid}",
  "type": "resume",
  "filename": "{filename}",
  "path": "{full path}",
  "target_company": null,
  "target_role": null,
  "date_created": "{file date or today}",
  "source": "existing",
  "notes": "Source document added to ExperienceLibrary"
}

6. Confirm

Added to ExperienceLibrary: {n} experience unit(s) from {filename}
  {Role} at {Company} ({dates})
  {Role} at {Company} ({dates})
  ...

ExperienceLibrary total: {total unit count} units

If duplicates were merged or skipped, note that inline. Keep the confirmation concise.