/setup

Auto-generated from skills/setup/SKILL.md

Academic Writer Setup

Quick onboarding wizard. Creates the researcher profile, detects integrations, and optionally fingerprints writing style. For deeper initialization (full 25-dimension style analysis, source indexing), run /academic-writer:init.

Phase 0: Preflight

Run silently before anything else:

mkdir -p past-articles .academic-helper .cognetivy/runs .cognetivy/events

If cognetivy is available, start a setup run:

echo '{"phase": "setup"}' > /tmp/aw-setup-input.json
cognetivy run start --input /tmp/aw-setup-input.json --name "Academic Writer Setup"

Capture the run_id for logging at each step.

Migrate any legacy profile from .academic-writer/profile.json → .academic-helper/profile.md:

python3 << 'PYTHON'
import os, json, re
from datetime import datetime

old_path = '.academic-writer/profile.json'
new_path = '.academic-helper/profile.md'
if os.path.exists(old_path) and not os.path.exists(new_path):
    with open(old_path) as f:
        p = json.load(f)
    scalar_keys = ['fieldOfStudy', 'citationStyle', 'targetLanguage', 'updatedAt', 'createdAt']
    list_keys = ['abstractLanguages', 'analyzedArticles']
    json_sections = [
        ('tools', 'Tools'), ('outputFormatPreferences', 'Output Format Preferences'),
        ('styleFingerprint', 'Style Fingerprint'), ('articleStructure', 'Article Structure'),
        ('sources', 'Sources'),
    ]
    lines = ['# Academic Writer Profile', '', '---']
    for k in scalar_keys:
        if k in p and p[k] is not None:
            lines.append(f'{k}: {p[k]}')
    for k in list_keys:
        v = p.get(k) or []
        if not v:
            lines.append(f'{k}: []')
        else:
            lines.append(f'{k}:')
            for item in v:
                lines.append(f'  - {item}')
    lines.append('---')
    lines.append('')
    for k, heading in json_sections:
        if k in p and p[k] is not None:
            lines.extend([f'## {heading}', '', '```json',
                          json.dumps(p[k], indent=2, ensure_ascii=False), '```', ''])
    os.makedirs('.academic-helper', exist_ok=True)
    with open(new_path, 'w') as f:
        f.write('\n'.join(lines))
    print(f"Migrated profile to {new_path}")
PYTHON

Check for existing profile:

cat .academic-helper/profile.md 2>/dev/null && echo "EXISTS" || echo "NOT_FOUND"

If EXISTS, load and show current settings, then:

AskUserQuestion(questions=[{
  "question": "A profile already exists. What would you like to do?",
  "header": "Existing profile detected",
  "options": [
    {
      "label": "Update existing profile",
      "description": "Keep existing data, modify only selected fields.",
      "markdown": "```\nUpdate Mode\n───────────\n✓ Existing settings preserved\n→ Only re-run steps you choose\n```"
    },
    {
      "label": "Start fresh",
      "description": "Delete and recreate from scratch.",
      "markdown": "```\nFresh Start\n───────────\n⚠ Current profile will be replaced\n→ Walk through all steps again\n```"
    }
  ],
  "multiSelect": false
}])

Phase 1: Detect Integrations

Run ALL detection commands in one parallel batch:

# PARALLEL
Bash(command="command -v ck >/dev/null 2>&1 && echo 'ck: DETECTED' || echo 'ck: NOT_FOUND'")
Bash(command="curl -s --max-time 3 http://localhost:8000/health 2>/dev/null && echo 'vectorless: RUNNING' || echo 'vectorless: NOT_RUNNING'")
Bash(command="command -v cognetivy >/dev/null 2>&1 && echo 'cognetivy: DETECTED' || echo 'cognetivy: NOT_FOUND'")
Bash(command="command -v nlm >/dev/null 2>&1 && nlm login --check 2>/dev/null && echo 'notebooklm: DETECTED' || echo 'notebooklm: NOT_FOUND'")

If vectorless NOT_RUNNING on port 8000:

AskUserQuestion(questions=[{
  "question": "Agentic-Search-Vectorless didn't respond on port 8000. What port is it running on?",
  "header": "Vectorless port",
  "options": [
    {"label": "Skip — not running right now", "description": "You can enable it later with /academic-writer:update-tools."}
  ],
  "multiSelect": false
}])

Retry with the provided port; save to tools.agentic-search-vectorless.port.

MongoDB Agent Skills is auto-configured silently — do not show it to the user.

Phase 2: Profile Setup

Field of Study

AskUserQuestion(questions=[{
  "question": "What is your field of study and area of specialization?",
  "header": "Step 1 — Field of Study",
  "options": []
}])

"The more specific, the better. Examples: Early Modern Jewish Philosophy, Talmudic Literature, Biblical Studies — Pentateuch"

Article Language

AskUserQuestion(questions=[{
  "question": "What language will you write your articles in?",
  "header": "Step 2 — Article Language",
  "options": [
    {
      "label": "Hebrew",
      "description": "RTL, David font, inline-parenthetical citations.",
      "markdown": "```\nHebrew Mode\n───────────\nDirection:  RTL\nFont:       David 11pt\nCitations:  (מחבר, כותרת, עמ' N)\n```"
    },
    {
      "label": "English",
      "description": "LTR, Times New Roman, Chicago/MLA/APA.",
      "markdown": "```\nEnglish Mode\n────────────\nDirection:  LTR\nFont:       Times New Roman 12pt\n```"
    },
    {
      "label": "Other",
      "description": "You'll type the language name in the next prompt.",
      "markdown": "```\nOther\n─────\n→ Type the language name\n→ RTL/LTR auto-detected\n```"
    }
  ],
  "multiSelect": false
}])

Citation Style

AskUserQuestion(questions=[{
  "question": "Which citation style do you use?",
  "header": "Step 3 — Citation Style",
  "options": [
    {
      "label": "Inline Parenthetical (Recommended for Hebrew)",
      "description": "(Author, Title, Page) in running text.",
      "markdown": "```\nExample: (לוי, משנת הנפש, עמ' 42)\n```"
    },
    {
      "label": "Chicago / Turabian",
      "description": "Footnotes with full bibliography.",
      "markdown": "```\nExample footnote: ¹ Levy, Soul's Teaching, 42.\n```"
    },
    {
      "label": "MLA",
      "description": "Parenthetical with Works Cited.",
      "markdown": "```\nExample: (Levy 42)\n```"
    },
    {
      "label": "APA",
      "description": "(Author, Year) — more common in social sciences.",
      "markdown": "```\nExample: (Levy, 2019, p. 42)\n```"
    }
  ],
  "multiSelect": false
}])

Tool Selection

AskUserQuestion(questions=[{
  "question": "Which integrations would you like to enable?",
  "header": "Step 4 — Tools",
  "options": [
    {
      "label": "Candlekeep",
      "description": "✓ Detected  /  ✗ Not found",
      "markdown": "```\nCandlekeep\n──────────\nType:  CLI (ck)\nWhat:  Cloud document library\n```"
    },
    {
      "label": "Agentic-Search-Vectorless",
      "description": "✓ Running  /  ✗ Not running",
      "markdown": "```\nAgentic-Search-Vectorless\n─────────────────────────\nType:  Local HTTP service\nWhat:  Fast semantic citation search\n```"
    },
    {
      "label": "Cognetivy",
      "description": "✓ Detected  /  ✗ Not found",
      "markdown": "```\nCognetivy\n─────────\nType:  CLI\nWhat:  Workflow audit trail\n\nInstall: npm install -g cognetivy\nInit:    timeout 5 cognetivy init --workspace-only\n```"
    },
    {
      "label": "NotebookLM",
      "description": "✓ Detected  /  ✗ Not found",
      "markdown": "```\nNotebookLM\n──────────\nType:  MCP server (nlm CLI)\nWhat:  AI-powered source Q&A,\n       audio overviews, study guides\n\nInstall: npm install -g notebooklm-mcp-cli\nAuth:    nlm login\n```"
    }
  ],
  "multiSelect": true
}])

Pre-check tools that were detected. MongoDB Agent Skills is silently included always.

Phase 3: Style Fingerprint (Optional)

Check for past articles:

ls past-articles/ 2>/dev/null | wc -l

AskUserQuestion(questions=[{
  "question": "Found N papers in past-articles/. Analyze them for style fingerprinting?",
  "header": "Writing style (optional)",
  "options": [
    {
      "label": "Yes, analyze my writing style (Recommended)",
      "description": "Extracts your voice across 25 dimensions so articles sound like you.",
      "markdown": "```\nStyle Analysis\n──────────────\n→ Reads PDF and DOCX in past-articles/\n→ Analyzes 25 dimensions:\n   sentence patterns, vocabulary,\n   paragraph structure, transitions,\n   citation density, rhetorical moves\n→ Shows fingerprint before saving\n→ You can correct anything\n```"
    },
    {
      "label": "Skip for now",
      "description": "Articles will use generic academic style until you run this.",
      "markdown": "```\nSkip\n────\n⚠ No style fingerprint yet\n→ Add papers to past-articles/ anytime\n→ Re-run: /academic-writer:init\n```"
    }
  ],
  "multiSelect": false
}])

If no files found, show only the "Skip" option with instructions to add papers.

If "Yes": analyze across all 25 dimensions (sentence level, vocabulary, paragraph structure, tone, transitions, citation style, rhetorical patterns, representative excerpts, article structure). Show fingerprint summary and confirm before saving.

Phase 4: Save Profile

Use the Write tool to create .academic-helper/profile.md:

# Academic Writer Profile

fieldOfStudy: FIELD
citationStyle: inline-parenthetical
targetLanguage: Hebrew
abstractLanguages:
  - Hebrew
analyzedArticles: []
createdAt: ISO_TIMESTAMP
updatedAt: ISO_TIMESTAMP

## Tools

```json
{
  "candlekeep": { "enabled": true },
  "agentic-search-vectorless": { "enabled": true, "port": 8000 },
  "cognetivy": { "enabled": true },
  "notebooklm": { "enabled": false }
}
` ` `

## Output Format Preferences

```json
{
  "font": "David",
  "bodySize": 11,
  "titleSize": 16,
  "headingSize": 13,
  "lineSpacing": 1.5,
  "marginInches": 1.0,
  "alignment": "justify",
  "rtl": true
}
` ` `

## Style Fingerprint

```json
null
` ` `

## Sources

```json
[]
` ` `

(Replace backtick triplets with actual ``` when writing the file.)

If Cognetivy is enabled, register workflows:

# Resolve plugin directory from Claude plugin cache
AW_PLUGIN_DIR=$(find ~/.claude/plugins/cache -name "wf_write_article.json" -path "*/workflows/*" 2>/dev/null | head -1 | xargs dirname 2>/dev/null)
if [ -z "$AW_PLUGIN_DIR" ]; then
  echo "Warning: Could not find academic-writer workflows in plugin cache. Skipping workflow registration."
else
  cognetivy workflow set --file "$AW_PLUGIN_DIR/wf_write_article.json"
  cognetivy workflow set --file "$AW_PLUGIN_DIR/wf_edit_article.json"
  cognetivy workflow set --file "$AW_PLUGIN_DIR/wf_edit_section.json"
  cognetivy workflow set --file "$AW_PLUGIN_DIR/wf_research.json"
  cognetivy workflow set --file "$AW_PLUGIN_DIR/wf_setup.json"
  cognetivy collection-schema set --file "$AW_PLUGIN_DIR/collection-schemas.json"
fi

Complete the Cognetivy run:

cognetivy run complete --run <run_id>

Completion

AskUserQuestion(questions=[{
  "question": "Setup complete. What would you like to do next?",
  "header": "You're all set!",
  "options": [
    {
      "label": "Write my first article → /academic-writer:write",
      "description": "Start the writing pipeline now.",
      "markdown": "```\n/academic-writer:write\n────────────────\nConversational pipeline:\nsubject → sources → thesis\n→ outline → write → .docx\n```"
    },
    {
      "label": "Run deeper initialization → /academic-writer:init",
      "description": "Full style analysis (25 dimensions) and source indexing.",
      "markdown": "```\n/academic-writer:init\n─────────────────────\n→ Deep 25-dimension fingerprint\n→ Article structure analysis\n→ Candlekeep source indexing\n```"
    },
    {
      "label": "Done for now",
      "description": "Profile saved. Run /academic-writer:write anytime.",
      "markdown": "```\nKey commands:\n  /academic-writer:write         ← write\n  /academic-writer:health  ← check\n  /academic-writer:update-tools ← change tools\n```"
    }
  ],
  "multiSelect": false
}])

Show summary table:

Setting	Value
Field	[field]
Language	[language]
Citation style	[style]
Tools enabled	[list]
Style fingerprint	✓ Analyzed / — Skipped