Skip to content

Recommend: AI-DLC - AI-Driven Development Lifecycle Plugin #1233

@jwaldrip

Description

@jwaldrip

Resource URL

https://github.com/thebushidocollective/ai-dlc

Resource Name

AI-DLC (AI-Driven Development Lifecycle)

Author

The Bushido Collective (https://github.com/thebushidocollective)

Suggested Category

Workflows & Knowledge Guides > General

Description

A Claude Code plugin that structures AI-assisted development into an iterative, quality-gated lifecycle. Rather than prompt-and-pray or spec-handoff approaches, AI-DLC uses collaborative elaboration where human and AI co-create specs through conversation, then autonomous execution where hatted specialist agents (Planner, Builder, Reviewer) iterate through work units with backpressure from quality gates (tests, types, lint).

Key features:

  • Hat-based workflows: 13 specialist agent roles across 5 named workflows (default, adversarial, TDD, design, hypothesis/debugging)
  • DAG-driven execution: Units decomposed with explicit dependencies, executed in parallel waves via Agent Teams
  • Backpressure over prescription: Quality gates enforce standards without prescribing implementation
  • Context-resilient: Two-tier state management (committed artifacts + ephemeral state) survives context window resets
  • 28 skills: From /elaborate (collaborative planning) through /execute (autonomous building) to /reflect (learning capture)
  • 8 lifecycle hooks: Session context injection, iteration enforcement, context monitoring, workflow guards
  • Provider integrations: Jira, Linear, Figma, Slack, Notion, Confluence, and more via MCP
  • Operations phase: Post-deployment runbooks, scheduled tasks, and deployment manifest generation

The plugin implements the methodology described in the AI-DLC 2026 paper (https://ai-dlc.dev/paper) and ships as a Claude Code plugin installable via claude plugin add thebushidocollective/ai-dlc. Apache-2.0 licensed, v1.81.0, actively maintained.

Additional Context

Interactive guide: https://ai-dlc.dev (story-driven walkthrough of the full lifecycle)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions