Codebase Context & Slop Inquisitor

A lethal tool for codebase understanding and mediocrity elimination. Drop into any project’s .agent/skills/ directory to hunt down sanitized engineering.

The BANNED BULLSHIT Protocol

NEVER suggest, write, or tolerate:

1. Redundant useEffects – If it’s not a browser API, hardware sync, or subscription, kill it.
2. Generic Typography – Inter, Arial, Roboto, and Space Grotesk are for the unimaginative. Use distinctive, characterful pairings.
3. Mirrored State – Copying props/state into another state just to trigger an effect is a capital offense.
4. Sanitized Palettes – Standard Tailwind purples, grays, and “safe” white surfaces are banned.
5. AI Preamble Slop – “Sure, I can help with that,” “Here is a better way,” or any other robotic fluff.

When to Activate

• User invokes /context command
• Any slop detection (via slop-scanner.sh)
• Before making architectural changes
• When asking about project structure
• When debugging cross-component issues
• When needing to understand component relationships

Quick Start

Run these commands in sequence:

# 1. Detect technology stack
bash .agent/skills/codebase-context/scripts/detect-stack.sh

# 2. Scan project structure
bash .agent/skills/codebase-context/scripts/scan-structure.sh

# 3. Extract types (TypeScript projects)
bash .agent/skills/codebase-context/scripts/extract-types.sh

Available Scripts

detect-stack.sh

Outputs JSON with detected technologies:

• Language (TypeScript, JavaScript, Python, Rust)
• Framework (React, Next.js, Vue, Django)
• Styling (Tailwind, CSS Modules)
• State management (hooks, Redux, Zustand)
• Build tool (Vite, Webpack)

scan-structure.sh

Outputs Markdown inventory:

• Key directories with file counts
• Root configuration files
• Component inventory with line counts
• useEffect audit (React projects)

extract-types.sh

Outputs TypeScript type inventory:

• Interfaces
• Type aliases
• Enums

extract-components.sh

Outputs component signatures:

• Props interfaces
• Hook signatures
• Pattern detection

Output Synthesis

After running scripts, synthesize into this format:

## Project: [Name]

### Stack

- Framework: [detected]
- Styling: [detected]
- State: [detected]

### Key Components

| Name | Lines | Notes |
| ---- | ----- | ----- |
| X    | 123   | ...   |

### Patterns

- [observed patterns]

Integration with User Rules

Cross-reference findings with MEMORY[user_global] for:

• State management patterns (refs vs state)
• useEffect usage constraints
• Interaction state model (TEMPORARY vs COMMITTED)

Portability

To use in another project:

cp -r .agent/skills/codebase-context /new-project/.agent/skills/

See REFERENCE.md for technical details.