Main Menu
  • Tools
  • Developers
  • Topics
  • Discussions
  • Communities
  • News
  • Podcasts
  • Blogs
  • Builds
  • Contests
  • Compare
  • Arena
Create
    EveryDev.ai
    Sign inSubscribe
    Home
    Tools

    2,342+ AI tools

    • New
    • Trending
    • Featured
    • Compare
    • Arena
    Categories
    • Agents1228
    • Coding1045
    • Infrastructure455
    • Marketing414
    • Design374
    • Projects340
    • Analytics319
    • Research306
    • Testing200
    • Data171
    • Integration169
    • Security169
    • MCP164
    • Learning146
    • Communication131
    • Prompts122
    • Extensions120
    • Commerce116
    • Voice107
    • DevOps92
    • Web73
    • Finance19
    1. Home
    2. Tools
    3. Draft CLI Plugin
    Draft CLI Plugin icon

    Draft CLI Plugin

    AI Coding Assistants
    Featured

    Persistent shared context plugin for Claude Code, Codex CLI, and Cursor that injects your full product context into every AI session automatically.

    Visit Website

    At a Glance

    Pricing
    Open Source

    Fully free and open-source plugin available on GitHub.

    Engagement

    Available On

    Windows
    macOS
    Linux
    API
    CLI

    Resources

    WebsiteGitHubllms.txt

    Topics

    AI Coding AssistantsAgent MemoryContext Engineering

    Alternatives

    XHawkQoderContextPool
    Developer
    idodekeroboidodekerobo builds open-source developer tools focused on AI…

    Listed May 2026

    About Draft CLI Plugin

    Draft is an open-source CLI plugin built by idodekerobo that solves the "blank slate" problem in AI coding assistants. It installs into Claude Code, Codex CLI, and Cursor, then automatically injects a live snapshot of your product context — company, product, team, priorities, and memory — before you type a single word in a new session. The project is hosted on GitHub, currently at version v2.2.1, and supports macOS and Linux only.

    What It Is

    Draft is a persistent context layer for AI-assisted development and product management workflows. Rather than re-explaining your product, stack, and priorities at the start of every Claude Code, Codex, or Cursor session, Draft's session hook fires automatically and loads a structured workspace snapshot. It targets the three failure modes the README identifies: context amnesia (blank slate on every session), context rot (stale docs diverging from reality), and context isolation (each teammate's AI starting cold).

    Agent Architecture

    Draft uses an orchestrator-plus-three-sub-agents pattern:

    • draft-agent — the main orchestrator; manages shared context and delegates to specialists
    • draft-researcher — reads workspace files and fetches web content when you need to know something
    • draft-executor — writes PRDs, decision docs, and updates files when you need to do something
    • draft-learner — updates context files and logs decisions when you need to remember something

    Most requests flow through researcher → executor → learner. The /draft:learn command (or $draft-learn on Codex) can also be invoked directly to capture a decision, scope change, or vocabulary term outside of a full PM session.

    Team Collaboration Model

    Draft supports sharing context across a team via a shared private GitHub repo. One person — the curator — runs /draft:setup-collab to configure the repo and /publish-team to push updates. Teammates run /load-team to pull the latest context. The shared layer covers context/ (company, product, team, priorities, decisions), while personal/ (working style, AI learnings, WIP drafts) always stays local. The README describes the intended feeling as: "every member of your team has an AI that was in the same room as everyone else."

    Multi-Editor Setup and Workspace Layout

    All three editors share a single workspace at ~/.draft/workspace/, which lives outside ~/.claude/, ~/.codex/, and ~/.cursor/ intentionally. Installing Draft for multiple editors does not create duplicate workspaces — each editor's setup script detects what's already installed and skips redundant rules or agent files. The recommended install order when using all three is Claude Code first, then Codex, then Cursor.

    Context is injected via editor-specific hook mechanisms:

    • Claude Code: SessionStart → inject-context.sh (raw text)
    • Codex: SessionStart → inject-context.sh (raw text as developer context)
    • Cursor: sessionStart → cursor-session-start.sh (JSON additional_context into initial system context)

    Update: v2.2.1

    The latest release is v2.2.1, published on 2026-05-14. The repository was created in April 2026 and last updated in May 2026, indicating active early development. The project uses a tagged release workflow on GitHub, with version tracking stored at ~/.draft/version and background update checks cached per session. The update command (/draft:update, $draft-update, or /draft-update depending on editor) shows the current and available versions, confirms before applying, and never touches context data at ~/.draft/workspace/.

    Draft CLI Plugin - 1

    Community Discussions

    Be the first to start a conversation about Draft CLI Plugin

    Share your experience with Draft CLI Plugin, ask questions, or help others learn from your insights.

    Pricing

    OPEN SOURCE

    Open Source

    Fully free and open-source plugin available on GitHub.

    • Persistent context injection for Claude Code, Codex CLI, and Cursor
    • Orchestrator + sub-agents pattern
    • Team collaboration via shared GitHub repo
    • Automatic background update checks
    • Multi-editor shared workspace

    Capabilities

    Key Features

    • Persistent context injection on every session start via SessionStart hooks
    • Orchestrator + three sub-agents pattern (draft-agent, draft-researcher, draft-executor, draft-learner)
    • Append-only decision log and context index to prevent context rot
    • Team collaboration via shared private GitHub repo (publish-team / load-team)
    • Multi-editor support: Claude Code, Codex CLI, and Cursor
    • Shared workspace at ~/.draft/workspace/ across all editors
    • /draft:learn command for manual capture of decisions, priorities, and vocabulary
    • Automatic background update checks with version tracking
    • Smart Cursor setup that detects existing Claude Code/Codex installs to avoid duplication
    • Personal layer (personal/) that is never shared with teammates
    • Context staleness policy (7-day / 21-day) enforced by draft-agent
    • Onboarding detection: auto-starts setup interview if no context exists

    Integrations

    Claude Code
    Codex CLI
    Cursor
    GitHub (shared private repo for team collaboration)
    gh CLI (for team collaboration setup)
    API Available

    Reviews & Ratings

    No ratings yet

    Be the first to rate Draft CLI Plugin and help others make informed decisions.

    Developer

    idodekerobo

    idodekerobo builds open-source developer tools focused on AI workflow enhancement. The Draft CLI plugin provides persistent shared context for Claude Code, Codex CLI, and Cursor, solving the blank-slate problem that affects AI-assisted development sessions. The project is hosted on GitHub and written primarily in Shell.

    Read more about idodekerobo
    WebsiteGitHub
    1 tool in directory

    Similar Tools

    XHawk icon

    XHawk

    Context infrastructure for AI-native development that automatically captures coding sessions, commits, and decisions into a living, searchable system of engineering knowledge.

    Qoder icon

    Qoder

    Agentic coding platform and AI-native IDE that uses context-aware agents to plan, code, test, and maintain software across desktop, IDE plugins, and CLI workflows.

    ContextPool icon

    ContextPool

    ContextPool gives AI coding agents persistent memory across sessions by extracting and storing engineering insights from past sessions via MCP.

    Browse all tools

    Related Topics

    AI Coding Assistants

    AI tools that help write, edit, and understand code with intelligent suggestions.

    445 tools

    Agent Memory

    Memory layers, frameworks, and services that enable AI agents to store, recall, and manage information across sessions. These tools provide persistent, semantic, and contextual memory for agents, supporting personalization, long-term context retention, graph-based relationships, and hybrid RAG + memory workflows.

    62 tools

    Context Engineering

    Techniques for optimizing context windows to improve AI responses.

    38 tools
    Browse all topics
    Back to all tools
    Explore AI Tools
    • AI Coding Assistants
    • Agent Frameworks
    • MCP Servers
    • AI Prompt Tools
    • Vibe Coding Tools
    • AI Design Tools
    • AI Database Tools
    • AI Website Builders
    • AI Testing Tools
    • LLM Evaluations
    Follow Us
    • X / Twitter
    • LinkedIn
    • Reddit
    • Discord
    • Threads
    • Bluesky
    • Mastodon
    • YouTube
    • GitHub
    • Instagram
    Get Started
    • About
    • Editorial Standards
    • Corrections & Disclosures
    • Community Guidelines
    • Advertise
    • Contact Us
    • Newsletter
    • Submit a Tool
    • Start a Discussion
    • Write A Blog
    • Share A Build
    • Terms of Service
    • Privacy Policy
    Explore with AI
    • ChatGPT
    • Gemini
    • Claude
    • Grok
    • Perplexity
    Agent Experience
    • llms.txt
    Theme
    With AI, Everyone is a Dev. EveryDev.ai © 2026
    Discussions