@jellylabs/didp

DIDP (Deterministic Iterative Development Protocol) skill for Claude Code.

Installation

Via npm (recommended)

# Install to current project
npx @jellylabs/didp install

# Install globally
npx @jellylabs/didp install --global

Via curl

# Install to current project
curl -sL https://jellylabs.ai/skills/didp-master/scripts/install.sh | bash

# Install globally
curl -sL https://jellylabs.ai/skills/didp-master/scripts/install.sh | bash -s global

Usage

After installation, use these slash commands in Claude Code:

| Command | Description | |---------|-------------| | /didp-init | Initialize DIDP in current project | | /didp-doctor | Run anti-pattern analysis | | /didp-doctor --diagnose | Check installation health | | /didp-update | Migrate from legacy DIDP | | /didp-start | Start new iteration |

CLI Commands

npx @jellylabs/didp install     # Install skill
npx @jellylabs/didp uninstall   # Remove skill
npx @jellylabs/didp doctor      # Run analysis
npx @jellylabs/didp diagnose    # Check installation
npx @jellylabs/didp help        # Show help

What is DIDP?

DIDP is a deterministic, phase-based workflow designed to make software development restartable, auditable, and safe under context compaction when using AI agents.

Core Principles

• Sessions are disposable; repository state is authoritative
• One iteration per session
• Phase-based workflow (planning → analysis → spec_lock → implementation → testing → complete)

Key Features

• Iteration state tracking via iteration_state.yaml
• Anti-pattern detection in CLAUDE.md files
• Automatic migration from legacy DIDP structures
• Installation diagnostics

Requirements

• Claude Code
• Bun runtime (for scripts)

Documentation

• Spec: https://jellylabs.ai/spec/didp/
• Methodology: https://jellylabs.ai/docs/methodology/
• Troubleshooting: See SKILL.md after installation

License

MIT