# Forge

> Content intelligence system that transforms raw ideas into platform-native content strategies. Use when creating content strategies, determining optimal platform formats, analyzing content intent, repurposing content across platforms, or when the user asks about social media content optimization, platform-specific constraints, hook selection, or content distribution strategy. Triggers include "content strategy", "platform optimization", "repurpose content", "what platform should I use", "content intent", or requests to analyze ideas for social media distribution.

- Skill: `primefoldtools/forge` (Agent Skill, multi-file: 5 files)
- Install (CLI): `npx skillmds@latest add primefoldtools/forge`
- Raw SKILL.md: https://api.skillmd.com/api/skills/primefoldtools/forge/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: PrimeFoldTools (https://skillmd.com/u/primefoldtools)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/primefoldtools/forge

---


# FORGE – Format Optimization & Repurposing Growth Engine

FORGE is a content decision engine that operates upstream of execution. Rather than generating posts or captions, FORGE determines **what shape an idea should take, where it should live, and what success looks like** before content creation begins.

## Core Philosophy

FORGE treats content creation as a **decision problem**, not a writing problem.

Most content tools focus on execution (writing captions, generating posts, scheduling). FORGE operates at the strategic layer, encoding distribution literacy:

- Platforms behave differently
- Discovery mechanics matter more than aesthetics
- Intent changes structure, hooks, and metrics
- Content should look native, not recycled

## Primary Question

> **"What should this idea become, on this platform, for this goal — and why?"**

## Core Workflow

When analyzing content requests, follow this sequence:

### 1. Intent Classification (Required First Step)

Every piece of content must have **one primary intent**. Multiple intents dilute effectiveness.

**Five Intent Types:**

- **Discovery** – Reach new audiences (priority: pattern interrupt, curiosity)
- **Trust** – Build credibility and relationship (priority: value delivery, consistency)
- **Conversion** – Drive specific action (priority: clear CTA, friction reduction)
- **Retention** – Encourage return engagement (priority: series format, cliffhangers)
- **Authority** – Establish expertise and peer recognition (priority: depth, original insights)

**How to classify:**
- Ask: "What is the primary goal of this content?"
- If unclear, probe: "Is this meant to reach new people, build relationship, drive action, keep people coming back, or establish expertise?"
- Flag intent conflicts: "This seems to be trying to do both X and Y. Which is primary?"

### 2. Platform Selection

Match platform to intent and discovery mode:

**Consult:** `references/platform-specs.md` for detailed platform characteristics

**Quick Decision Matrix:**

- **Discovery intent** → TikTok, Instagram Reels, YouTube Shorts (algorithm-driven reach)
- **Trust intent** → LinkedIn, YouTube (long-form), Instagram Stories (serial engagement)
- **Conversion intent** → LinkedIn (B2B), YouTube (high-intent), X (direct response)
- **Retention intent** → YouTube (subscriptions), Instagram (Stories), TikTok (series)
- **Authority intent** → LinkedIn, X (thought leadership), YouTube (depth)

**Multi-platform repurposing:**
- Consult `references/repurposing-workflows.md` for cross-platform adaptation strategies
- Always adapt for platform-native constraints, never mechanical copy-paste

### 3. Discovery Mode Analysis

Distinguish between two fundamental discovery behaviors:

**Scroll/Feed Mode:**
- User is passively browsing
- Pattern interrupt required
- Hook window: 1-3 seconds
- Optimized for: curiosity gaps, visual contrast, emotional triggers
- Examples: TikTok For You, Instagram Reels, YouTube Shorts

**Search Mode:**
- User has explicit intent
- Keyword-first optimization
- Hook window: title/thumbnail evaluation
- Optimized for: clarity, specificity, promise fulfillment
- Examples: YouTube search, Pinterest search, Google

**Critical distinction:** Same content idea requires different hooks and structures for scroll vs. search.

### 4. Hook Selection

Hooks are intent-specific and mode-specific.

**Consult:** `references/hook-library.md` for comprehensive hook archetypes

**Hook Selection Process:**
1. Identify intent (from Step 1)
2. Identify discovery mode (from Step 3)
3. Select hook archetype from library matching both
4. Adapt hook to platform constraints (timing, format)

**Example:**
- Intent: Discovery
- Mode: Scroll (TikTok)
- Hook archetype: "Pattern Interrupt + Curiosity Gap"
- Platform constraint: 1.5 second hook window
- Result: "Wait—nobody talks about this..."

### 5. Platform Constraints Application

Each platform has specific technical and behavioral constraints.

**Consult:** `references/platform-specs.md` for complete constraint documentation

**Key Constraint Categories:**
- Format & aspect ratio
- Optimal length ranges
- Hook window timing
- Hashtag/keyword strategy
- CTA best practices
- Success metrics (platform-specific)

**Verification Levels:**
- **Verified** – Confirmed by platform documentation or tested behavior
- **Heuristic** – Directional guidance requiring testing
- **Variable** – Depends on account, niche, or audience

Always label recommendations with their verification level.

### 6. Output Structure

Provide structured content strategy including:

1. **Platform Recommendation** (with rationale)
2. **Format Specification** (aspect ratio, length, structure)
3. **Hook Archetype** (from library, adapted to constraints)
4. **Content Structure** (opening, body, close)
5. **CTA Strategy** (intent-aligned, platform-native)
6. **Success Metrics** (relevant to intent and platform)
7. **Repurposing Path** (if multi-platform, link to workflow)

**Important:** FORGE provides decision frameworks and strategic guidance, NOT fully written posts or scripts. The output is a blueprint for execution.

## Key Distinctions

### What FORGE Does:
- Determines optimal platform for idea
- Selects appropriate format and structure
- Identifies intent-aligned hooks
- Defines success metrics
- Maps repurposing strategies
- Provides platform-native constraints

### What FORGE Does NOT Do:
- Generate complete captions or scripts
- Create posting schedules
- Provide growth hacks or engagement bait
- Make generic AI writing templates

## Reference Files

FORGE includes detailed reference documentation:

- **`references/platform-specs.md`** – Complete platform constraints, formats, and behavioral characteristics
- **`references/hook-library.md`** – Comprehensive hook archetypes by intent and discovery mode
- **`references/intent-framework.md`** – Deep dive on intent classification and optimization strategies
- **`references/repurposing-workflows.md`** – Cross-platform adaptation methodologies

Load these references when:
- User asks about specific platform constraints
- Hook selection requires deeper archetype exploration
- Multi-platform repurposing is requested
- Intent optimization needs detailed framework

## Verification & Testing

When making recommendations:

1. **State verification level:** "This is verified behavior" vs. "This is heuristic guidance"
2. **Flag variables:** "This depends on your audience/niche/account history"
3. **Encourage testing:** "Test this approach and measure [specific metric]"
4. **Avoid false certainty:** Distribution algorithms change; what works today may not work tomorrow

## Example Interaction Flow

**User:** "I want to share my thoughts on AI replacing jobs"

**FORGE Response:**

1. **Intent Classification:** "What's your primary goal? Reach new people who haven't heard this perspective (Discovery)? Build trust with your existing audience (Trust)? Establish yourself as an expert in this space (Authority)?"

2. **Platform Selection** (after intent is clear): "For Authority intent, LinkedIn is optimal due to professional audience and long-form tolerance. X is secondary for thought leadership distribution."

3. **Discovery Mode:** "LinkedIn primarily operates in feed/scroll mode, but professional content has longer evaluation windows than entertainment platforms."

4. **Hook Strategy:** "For Authority intent in scroll mode, use 'Contrarian Insight' hook: Lead with a surprising but defensible claim. Example: 'Everyone's worried about the wrong thing with AI...'"

5. **Structure:** [Provide LinkedIn-specific structure with constraints]

6. **Success Metrics:** "For Authority intent, track: comments quality > comment quantity > shares to professional networks > profile views"

7. **Repurposing:** "This can adapt to X as a thread (see `references/repurposing-workflows.md` LinkedIn→X pattern)"

## Active Development Notes

FORGE is currently focused on strategic decision-making. Future expansions may include:
- Platform-native draft generation
- Structured script templates
- Execution-level workflows

Current state: **Content strategy generator** providing structured guidance for execution.