/cf:help

Show the ContentForge user guide with setup instructions, pipeline overview, available skills, usage examples, and troubleshooting.

Behavior

When invoked, display a structured help overview. Use the comprehensive documentation in docs/USER-GUIDE.md for full details.

1. Quick Start Summary

Display this quick orientation:

=== CONTENTFORGE — HELP ===

Version: 3.8.0
Agents: 13 (10-phase pipeline + 3 post-pipeline)
Skills: 19 slash commands — all with argument-hint autocomplete
Connectors: 9 HTTP + 19 npx integrations (incl. image generation)
Tracking: 3 backends (Google Sheets + Drive, Airtable, Local)

Getting Started:
  1. /cf:style-guide         — Create your brand profile (start here)
  2. /cf:integrations         — See which connectors are active
  3. /contentforge            — Run the full 10-phase content pipeline
  4. /cf:help --examples      — See example prompts and workflows

2. Arguments

Argument	Effect
(none)	Show the full help overview
--pipeline	Show the 10-phase pipeline with timing and quality gates
--skills	List all 19 skills with descriptions
--brand	Explain brand profile setup methods
--examples	Show example workflows from brief to publish
--troubleshoot	Show common issues and solutions
--connectors	Show connector status (shortcut for /cf:integrations)

3. Pipeline Overview

When --pipeline is specified, show the 10-phase pipeline:

=== CONTENTFORGE PIPELINE ===

Step 0.5: Title Curation (1-2 min) — MANDATORY, runs before pipeline
  → SERP recon on top 5 results for your keyword
  → Generates 4-5 SEO-optimized title options (angles vary by content type)
  → Validates against brand guardrails, 60-char SERP limit, anti-clickbait
  → YOU select, modify, combine, or provide your own title
  → Pipeline only starts after you confirm a title

Phase 1: Researcher (3-5 min)
  → Uses your confirmed title as anchor for all research
  → Web research, 15-25 sources, relevance scoring

Phase 2: Fact Checker (2-3 min)
  → URL verification, claim validation, source grading

Phase 3: Content Drafter (5-8 min)
  → Brand-voiced draft with inline citations, SME calibration via industry knowledge packs

Phase 3.5: Visual Asset Annotator (1-2 min)
  → Chart generation from verified stats, visual markers, asset manifest

Phase 4: Scientific Validator (2-3 min)
  → Hallucination detection, domain-specific validation, claim cross-referencing

Phase 5: Structurer & Proofreader (2-3 min)
  → Heading hierarchy, readability, grammar, formatting

Phase 6: SEO/GEO Optimizer (2-3 min)
  → Keywords, meta tags, schema, AI Overview optimization, internal linking

Phase 6.5: Humanizer (2-3 min)
  → AI pattern removal, personality profiles, industry patterns

Phase 7: Reviewer (2-3 min)
  → 5-dimension scoring (needs ≥7.0), visual + link quality, comparative analysis

Phase 8: Output Manager (1-2 min)
  → .docx with embedded charts, internal links, backend-dispatched delivery + tracking
  → Pipeline performance report: actual timing per phase, token estimate, benchmarks

Post-Pipeline:
  → Social Adapter (#10): Article → social posts
  → Translator (#11): Multilingual with brand voice
  → Batch Orchestrator (#9): Parallel processing

Quality Gates: Score ≥7.0 required | Max 2 loops per phase | 5 total loops max
Three-Layer Verification: Fact Checker → Scientific Validator → Reviewer
Result: Zero hallucinations in production

4. All Skills

When --skills is specified, list all 19 skills:

Skill	Description
/contentforge	Run the full 10-phase content production pipeline
/batch-process	Process 10-50+ pieces in parallel from Google Sheets
/content-refresh	Update existing content with fresh data and sources
/cf:style-guide	Create or update a brand profile interactively
/cf:integrations	See which connectors are active and available
/cf:connect <name>	Step-by-step setup guide for any connector
/cf:add-integration	Add a custom MCP connector for any API or service
/cf:publish	Push content to Webflow/WordPress or export HTML
/cf:social-adapt	Transform article into social posts (5 platforms)
/cf:variants	Generate A/B test variations of headlines, hooks, CTAs
/cf:analytics	Quality score trends, timing breakdown, brand performance
/cf:translate	Translate preserving brand voice (15+ languages)
/cf:video-script	Timestamped video scripts for YouTube, TikTok, Reels
/cf:brief	Research-backed content brief with keyword analysis
/cf:audit	Content freshness scoring, decay detection, gap analysis
/cf:calendar	Production scheduling with deadline tracking
/cf:template	Create custom content type templates
/cf:switch-backend	Switch tracking backend (local/airtable/google) with migration
/cf:help	This help guide

5. Brand Setup Quick Reference

When --brand is specified, explain the 3 setup methods:

=== BRAND PROFILE SETUP ===

Method 1: Interactive (Recommended)
  /cf:style-guide
  → Answer questions about voice, terminology, industry, guardrails
  → Generates brand profile JSON automatically

Method 2: Manual JSON
  → Copy config/brand-registry-template.json
  → Fill in your brand details
  → Save to your brand config directory

Method 3: Google Drive Knowledge Vault
  → Create a folder: "ContentForge Brands/[Brand Name]/Knowledge Vault"
  → Add documents: voice-and-tone.md, terminology.md, prohibited-claims.md
  → ContentForge auto-extracts on first run (SHA256 cached)

Brand Profile Includes:
  - Voice & Tone (authoritative, conversational, technical, witty)
  - Terminology (approved terms, banned phrases)
  - Style Guide (formatting, citation style)
  - Guardrails (topics to avoid, compliance)
  - Industry Context (Pharma, BFSI, Healthcare, Legal)
  - Personality Profile (4 profiles for Humanizer)
  - Tracking Backend (Google Sheets + Drive, Airtable, or Local)

Profiles are cached (SHA256 hash) — 95% time savings on repeat runs.
Switch backends anytime: /cf:switch-backend [airtable|google|local]

6. Example Workflows

When --examples is specified, show practical end-to-end examples:

=== EXAMPLE: First Article ===

Step 1: Create brand profile
  /cf:style-guide
  → Provide: Brand name, industry, voice, terminology, guardrails

Step 2: Generate content brief (optional but recommended)
  /cf:brief "AI-Powered Diagnostics in Healthcare"
  → Gets: Keyword analysis, competitor review, recommended outline

Step 3: Run the pipeline
  /contentforge "AI-Powered Diagnostics in Healthcare"
    --type=article --brand=AcmeMed --audience="Healthcare Executives"
    --keyword="AI diagnostics precision medicine"
  → Title Curation: 4-5 options generated from SERP + brand voice
    1. "AI Diagnostics in Healthcare: What Clinicians Need to Know" (54 chars) — authority
    2. "How AI Is Transforming Medical Diagnostics in 2026" (50 chars) — trending
    3. "5 Ways AI Diagnostics Improve Patient Outcomes" (47 chars) — listicle
    4. "AI-Powered Diagnostics: The Future of Precision Medicine" (56 chars) — future-focused
    5. "Why Top Hospitals Are Betting on AI Diagnostics" (48 chars) — expert
  → You pick, modify, or write your own title
  → Pipeline runs with confirmed title → 20-30 min → Quality score 9.1/10

Step 4: Publish
  /cf:publish --platform=webflow
  → Or: /cf:publish --export=html (for manual upload)

Step 5: Create social posts
  /cf:social-adapt
  → LinkedIn, Twitter/X, Instagram, Facebook, Threads posts

Step 6: Translate for global audience
  /cf:translate --language=es --level=adapted
  → Spanish version preserving brand voice


=== EXAMPLE: Batch Processing ===

Step 1: Prepare Google Sheets with columns A-N:
  A: Requirement ID | B: Content Type | C: Title | D: Audience
  E: Brand | F: Primary Keyword | G: Target Words | H: Priority
  I: Tone Override | J: Status | K: Score | L: Output URL
  M: Processing Time | N: Completed At

Step 2: Run batch
  /batch-process --sheet="Content Q1 2026" --rows=1-20
  → Processes 20 pieces in parallel (4-5x faster)

Step 3: Review results
  /cf:analytics --period=30
  → Quality trends, timing breakdown, brand performance

7. Skill Platform Features

When showing the full help, include:

=== SKILL PLATFORM FEATURES ===

Argument Hints (16 skills):
  All skills show autocomplete hints in the Skills UI.
  Example: /contentforge shows "topic" --type=article --brand=name
  Example: /cf:brief shows "topic or keyword" [--depth=deep]

Execution Safety:
  /cf:publish requires explicit user invocation — Claude cannot
  auto-trigger it. Prevents accidental publishing to Webflow/WordPress.

Quality Evals (3 skills):
  contentforge, cf-brief, and cf-style-guide have evals/evals.json
  with structured test cases for quality benchmarking.

8. Troubleshooting

When --troubleshoot is specified, show common issues:

Issue	Solution
"Brand profile not found"	Run /cf:style-guide to create your brand profile
Quality score below 5.0	Content flagged for human review — check topic complexity and source availability
"Research timeout"	Check internet connection; ContentForge needs web access for Phase 1 research
Google Drive not showing in connectors	Google Drive is a platform-level integration — check Claude Desktop → Settings → Integrations
MCP connector not working	Run /cf:integrations to check status, /cf:connect <name> for setup
Pipeline taking too long	Normal: 20-30 min for articles. For faster: use /batch-process for parallel processing
Humanizer removing too much	Adjust personality profile: /contentforge --tone=conversational for lighter touch

9. Documentation References

Point users to these resources:

Guide	What it covers
docs/USER-GUIDE.md	Comprehensive end-to-end guide (1,300+ lines)
UPGRADE-GUIDE.md	v2.1.0 → v3.0.0 migration guide (historical)
CONNECTORS.md	All available connectors by category
CHANGELOG.md	Full version history and release notes
config/brand-registry-template.json	Brand profile JSON template
config/social-platform-specs.json	Social platform constraints and specs
config/multilingual-patterns.json	Language-specific brand voice patterns
templates/	Content type templates and formats

Output Format

Present information in clean, scannable tables and code blocks. Keep the output concise. Link to docs/USER-GUIDE.md for full walkthroughs.