devlyn-cli

Context Engineering & Harness Engineering Toolkit for Claude Code

Structured prompts, agent orchestration, and automated pipelines — debugging, code review, UI design, product specs, and more.

Get Started · Commands · Skills · Workflows · Optional Packs · Contributing

Why devlyn-cli?

Claude Code is powerful out of the box — but teams need consistent, repeatable workflows. Without shared conventions, every developer prompts differently, reviews differently, and debugs differently.

devlyn-cli solves this with two complementary engineering approaches:

Context Engineering

Structured prompts and role-based instructions that shape what the AI knows and how it thinks for each task.

• 16 slash commands for debugging, code review, UI design, documentation, and more
• 5 core skills that activate automatically based on conversation context
• Agent team workflows that spawn specialized AI teammates with role-specific expertise
• Product & feature spec templates for structured planning

Harness Engineering

Pipeline orchestration that controls how agents execute — permissions, state management, multi-phase workflows, and cross-model evaluation.

• /devlyn:auto-resolve — 9-phase automated pipeline (build → browser validate → evaluate → fix loop → simplify → review → security → clean → docs)
• /devlyn:browser-validate — feature verification in a real browser with tiered fallback (Chrome MCP → Playwright → curl)
• bypassPermissions mode for autonomous subagent execution
• File-based state machine — agents communicate via .claude/done-criteria.md, EVAL-FINDINGS.md, and BROWSER-RESULTS.md
• Git checkpoints at each phase for rollback safety
• Cross-model evaluation via --with-codex flag (OpenAI Codex as independent evaluator)

Zero dependencies. One command. Works with any project.

Get Started

npx devlyn-cli

The interactive installer walks you through setup — select optional skills, choose community packs, done.

# Non-interactive install (CI/CD friendly)
npx devlyn-cli -y

# Update to the latest version
npx devlyn-cli@latest

# See everything that's included
npx devlyn-cli list

What Gets Installed

your-project/
├── .claude/
│   ├── commands/              # 16 slash commands
│   ├── skills/                # 5 core skills + any optional addons
│   ├── templates/             # Product spec, feature spec, prompt templates
│   ├── commit-conventions.md  # Commit message standards
│   └── settings.json          # Agent teams enabled
└── CLAUDE.md                  # Project-level AI instructions

Commands

Slash commands are invoked directly in Claude Code conversations (e.g., type /devlyn:resolve).

Debugging & Resolution

Command	Description
/devlyn:resolve	Systematic bug fixing with root-cause analysis and test-driven validation
/devlyn:team-resolve	Spawns a full agent team — root cause analyst, test engineer, security auditor — to investigate complex issues
/devlyn:auto-resolve	Fully automated pipeline for any task — bugs, features, refactors, chores. Build → browser validate → evaluate → fix loop → simplify → review → clean → docs. One command, zero human intervention. Supports --with-codex for cross-model evaluation via OpenAI Codex
/devlyn:browser-validate	Verify implemented features work in a real browser — starts dev server, tests the feature end-to-end (clicks, forms, verification), with tiered fallback (Chrome MCP → Playwright → curl)

Code Review & Quality

Command	Description
/devlyn:review	Post-implementation review — security, quality, best practices checklist
/devlyn:team-review	Multi-perspective team review with specialized reviewers (security, quality, testing, performance, product)
/devlyn:evaluate	Independent quality evaluation — assembles evaluator team to grade work against done criteria with calibrated, skeptical grading
/devlyn:clean	Detect and remove dead code, unused dependencies, complexity hotspots, and tech debt

UI Design & Implementation

Command	Description
/devlyn:design-ui	Generate 5 radically distinct UI style explorations from a spec or reference image
/devlyn:team-design-ui	Spawns a design team — creative director, product designer, visual designer, interaction designer, accessibility designer
/devlyn:design-system	Extract design system tokens from a chosen style for exact reproduction
/devlyn:implement-ui	Team-based UI build — component architect, UX engineer, accessibility engineer, responsive engineer, visual QA

Product & Planning

Command	Description
/devlyn:product-spec	Generate or incrementally update product spec documents
/devlyn:feature-spec	Transform product specs into implementable feature specifications
/devlyn:discover-product	Scan codebase to generate feature-oriented product documentation
/devlyn:recommend-features	Prioritize top 5 features to build next based on value and readiness

Documentation

Command	Description
/devlyn:update-docs	Sync all project docs with current codebase — cleans stale content, preserves roadmaps, generates missing docs

Skills

Skills are not invoked manually — they activate automatically when Claude Code detects a relevant conversation context. Think of them as always-on expertise that shapes how the AI approaches specific types of work.

Skill	When It Activates	What It Does
root-cause-analysis	Debugging conversations	Enforces 5 Whys methodology, evidence standards, and no-workaround rules
code-review-standards	Code review tasks	Applies severity framework, quality bar, and approval criteria
ui-implementation-standards	UI/frontend work	Ensures design fidelity, accessibility, animation quality, and responsive standards
code-health-standards	Code maintenance	Enforces dead code prevention, dependency discipline, and complexity thresholds
workflow-routing	Any task	SDLC phase map — guides you to the right command for your current task

Workflows

Commands are designed to compose. Pick the right tool based on scope, then chain them together.

Automated Pipeline (Recommended)

One command runs the full cycle — no human intervention needed:

/devlyn:auto-resolve fix the auth bug where users see blank screen on 401

Phase	What Happens
Build	team-resolve investigates and implements, writes testable done criteria
Browser Validate	For web projects: starts dev server, tests the implemented feature end-to-end in a real browser, fixes issues found
Evaluate	Independent evaluator grades against done criteria with calibrated skepticism
Fix Loop	If evaluation fails, fixes findings and re-evaluates (up to N rounds)
Simplify	Quick cleanup pass for reuse and efficiency
Review	Multi-perspective team review
Security	Dedicated OWASP-focused audit (auto-detects when changes touch auth, secrets, APIs)
Clean	Remove dead code and unused dependencies
Docs	Sync documentation with changes

Each phase runs as a separate subagent (fresh context), communicates via files, and commits a git checkpoint for rollback safety. Skip phases with flags: --skip-browser, --skip-review, --skip-clean, --skip-docs, --max-rounds 3, --with-codex (cross-model evaluation via OpenAI Codex).

Manual Workflow

For step-by-step control between phases:

Step	Command	What It Does
1. Resolve	/devlyn:resolve or /devlyn:team-resolve	Fix the issue — solo for focused bugs (1-2 modules), team for complex issues (3+ modules)
2. Evaluate	/devlyn:evaluate	Independent quality evaluation — grades against done criteria written in step 1
		If the evaluation finds issues: /devlyn:team-resolve "Fix issues in .claude/EVAL-FINDINGS.md"
3. Simplify	/simplify	Quick cleanup pass for reuse, quality, and efficiency (built-in Claude Code command)
4. Review	/devlyn:review or /devlyn:team-review	Audit the changes — solo for small PRs (< 10 files), team for large PRs (10+ files)
5. Clean	/devlyn:clean	Remove dead code, unused dependencies, and complexity hotspots
6. Document	/devlyn:update-docs	Sync project documentation with the current codebase

Steps 5-6 are optional — run them periodically rather than on every PR.

Scope matching matters. For a simple one-file bug, /devlyn:resolve + /devlyn:review (solo) is fast. For a multi-module feature, /devlyn:auto-resolve handles everything. Don't over-tool simple changes.

UI Design Pipeline

A full explore → extract → build pipeline:

Step	Command	What It Does
1. Explore	/devlyn:design-ui	Generates 5 radically distinct style options from a spec or reference image
2. Extract	/devlyn:design-system	Pulls exact design tokens (colors, spacing, typography) from your chosen style
3. Build	/devlyn:implement-ui	Spawns a build team (component architect, UX engineer, accessibility engineer, responsive engineer, visual QA)

For design exploration with a full creative team, use /devlyn:team-design-ui instead of step 1.

After building, follow the recommended workflow starting from step 2 (simplify) to review and polish the implementation.

Standalone Tools

These commands work independently, outside of the main workflow:

Command	What It Does
/devlyn:clean [focus]	Focused cleanup — e.g., /devlyn:clean dependencies or /devlyn:clean dead-code
/devlyn:update-docs [area]	Focused doc sync — e.g., /devlyn:update-docs API reference
/devlyn:product-spec	Generate or update a product specification
/devlyn:feature-spec	Transform a product spec into an implementable feature spec
/devlyn:discover-product	Scan codebase to auto-generate product documentation
/devlyn:recommend-features	Prioritize top 5 features to build next

Optional Skills & Packs

During installation, the interactive selector lets you add optional skills and community packs.

Skills

Copied directly into your .claude/skills/ directory.

Skill	Description
cloudflare-nextjs-setup	Cloudflare Workers + Next.js deployment with OpenNext
generate-skill	Create well-structured Claude Code skills following Anthropic best practices
prompt-engineering	Claude 4 prompt optimization using official Anthropic best practices
better-auth-setup	Production-ready Better Auth + Hono + Drizzle + PostgreSQL auth setup
pyx-scan	Check whether an AI agent skill is safe before installing
dokkit	Document template filling for DOCX/HWPX — ingest, fill, review, export
devlyn:pencil-pull	Pull Pencil designs into code with exact visual fidelity
devlyn:pencil-push	Push codebase UI to Pencil canvas for design sync

Community Packs

Installed via the skills CLI (npx skills add). These are maintained by their respective communities.

Pack	Description
vercel-labs/agent-skills	React, Next.js, React Native best practices
supabase/agent-skills	Supabase integration patterns
coreyhaines31/marketingskills	Marketing automation and content skills
anthropics/skills	Official Anthropic skill-creator with eval framework and description optimizer
Leonxlnx/taste-skill	Premium frontend design skills — modern layouts, animations, and visual refinement

MCP Servers

Installed via claude mcp add during setup.

Server	Description
codex-cli	Codex MCP server for cross-model evaluation via OpenAI Codex
playwright	Playwright MCP for browser testing — powers devlyn:browser-validate Tier 2

Want to add a pack? Open a PR adding your pack to the OPTIONAL_ADDONS array in bin/devlyn.js.

How It Works

• Run npx devlyn-cli in your project root
• The CLI copies config files into .claude/ and CLAUDE.md to the project root
• Claude Code automatically reads .claude/commands/ and .claude/skills/ on startup
• Invoke commands like /devlyn:resolve in your Claude Code session — skills activate on their own

The installation is idempotent — run it again anytime to update to the latest config.

Creating Your Own Skills

Want to author custom skills for your team or the community?

• During install, select the generate-skill optional skill — or —
• Install the official Anthropic skill-creator pack:

  npx skills add anthropics/skills
  

Both provide structured templates, best practices, and eval frameworks for writing high-quality Claude Code skills.

See the Claude Code skills documentation for the full specification.

Requirements

• Node.js 18+
• Claude Code CLI installed and configured

Contributing

Contributions are welcome! Here are some ways to get involved:

• Add a command — Create a new .md file in config/commands/
• Add a skill — Create a new directory in config/skills/ with a SKILL.md
• Add an optional skill — Add to optional-skills/ and the OPTIONAL_ADDONS array
• Suggest a community pack — Open a PR to add it to the pack list

Development

• Fork the repository
• Create your branch (git checkout -b feat/my-feature)
• Commit your changes following the included commit conventions
• Push to the branch (git push origin feat/my-feature)
• Open a Pull Request

Star History

License

MIT — Donut Studio