Dl

devloop

A Claude Code plugin that structures AI-assisted development into phases. Without structure, coding agents generate inconsistent patterns and one-shot attempts that miss edge cases. devloop fixes this by breaking work into brainstorming, research, planning, design, implementation, and review — each phase produces an artifact the next one reads. No step touches code until /dl:implement.

The loop

  brainstorm → research → plan → design → implement → review
                  ↑          ↑       ↑         │
                  └──────────┴───────┴─────────┘
                   re-enter research when
                   discoveries surface

• brainstorm — required entry point; iterative questioning to refine a feature idea, producing decisions and research queries
• research — execute research queries from brainstorm as targeted codebase searches
• plan — ask clarifying questions, produce an ordered task list
• design — generate architecture, Mermaid diagrams, and a detailed spec for each task
• implement — implement one task at a time against the spec, track completion
• review — load rules and design context, review code for rule violations and security issues

Start with /dl:brainstorm — it's the required entry point. /dl:research can re-enter at any stage when discoveries surface during implementation.

Install

From marketplace

claude plugin marketplace add minusblindfold/devloop
claude plugin install dl

For development

git clone https://github.com/minusblindfold/devloop.git
claude --plugin-dir ./devloop

Use /reload-plugins after making changes during development.

Skills

Skill	What it does
/dl:brainstorm	Required entry point. Iterative questioning to refine a feature idea. Produces decisions and research queries.
/dl:research	Execute research queries from brainstorm as targeted codebase searches. Requires brainstorm artifact.
/dl:plan	Ask clarifying questions, create a vertically-sliced task list. Detects greenfield projects.
/dl:design	Primary review checkpoint. Generate architecture, Mermaid diagrams, and per-task specs.
/dl:implement	Implement one task at a time against the spec. Tracks progress across sessions.
/dl:review	Load rules and design context, review code for rule violations and security issues. Works standalone or after /dl:implement.
/dl:plan and /dl:design support refine mode — invoke with no args when existing artifacts are found.

What you get

• Architecture diagrams — /dl:design generates Mermaid diagrams (architecture, data flow, component, sequence) saved as .mmd files in .work/designs/diagrams/.
• Cross-session tracking — Task completion persists. Run /dl:implement in a new session and pick up where you left off.
• Refine mode — Run /dl:plan or /dl:design with no args to iterate on existing artifacts.
• Greenfield detection — In a new project with no existing structure, /dl:plan suggests scaffolding as the first task.
• Active feature tracking — Skills auto-detect the active feature via .work/active/<slug>.md markers. Start a brainstorm or plan, and every downstream skill auto-selects it — no re-specifying slugs. Supports multiple concurrent features for multi-agent workflows.
• Rule-driven consistency — Skills match task descriptions against rule keywords, so the same patterns apply everywhere.
• Linked repositories — Declare related repos in rule frontmatter and /dl:brainstorm and /dl:research will scan them for cross-repo context.

For a deeper walkthrough, see docs/workflow.md.

Rules

Rules are markdown files that guide skills at runtime — coding patterns, project structure, naming conventions. Skills resolve rules from four layers, highest precedence first:

Precedence	Layer	Path	Description
1 (highest)	User	~/.claude/rules/	Claude Code native, personal overrides
2	Project	{cwd}/devloop/rules/	Project-specific rules
3	Shared/org	~/devloop/rules/	Rule packs managed by devloop CLI
4 (lowest)	Plugin-bundled	${CLAUDE_PLUGIN_ROOT}/rules/	Defaults shipped with devloop

Drop .md files with YAML frontmatter into any of these directories and skills pick them up automatically via keyword matching. For project-specific rules, create a devloop/rules/ directory in your project root and commit it to version control.

---
keywords: [entity, model, JPA, persistence]
---
# Entity Rules

## Patterns

- Pattern one — what to do and why.

Rules can also declare linked repositories with repos — paths to other local repos that skills should scan for cross-repo context:

---
keywords: [ecosystem, integration]
repos: [~/code/payments-service, ~/code/shared-types]
---

When /dl:brainstorm or /dl:research resolve a rule with repos, they scan those directories for relevant code, API contracts, and shared types. Paths must be home-relative (~/...). See devloop/rules/ecosystem.md for a working example.

Rules are optional. Without them, skills work from codebase context alone. Start by dropping .md files with keyword frontmatter into ~/.claude/rules/ — skills discover them automatically. When you want organized, reusable rule sets, see devloop-rules for packs with a management CLI. The full format spec is in plugins/dl/rules/rules.md.

Working directory

Skills read and write artifacts to .work/ in the current project directory:

.work/
├── active/          # Active feature markers for workflow continuity
├── brainstorms/     # Feature brainstorm decision logs
├── research/        # Research artifacts
├── plans/           # Task lists
├── designs/         # Design docs + diagrams/
├── implementations/ # Implementation notes
└── reviews/         # Code review findings

Artifacts follow the naming convention YYYY-MM-DD-<slug>-<type>.md (e.g., 2026-03-11-auth-design.md). This keeps them sortable and identifiable across features. See docs/artifacts.md for details on how artifacts chain together.

Add .work/ to your .gitignore.

Terminal companion tools

For terminal-based artifact browsing (view-brainstorm, view-plan, view-design, view-research, view-implement, open-diagrams, claude-context), see devenv. These are optional CLI tools that complement the plugin.

License

MIT