primeline-skills

PrimeLine Skills

5 production-grade workflow skills for Claude Code.

Debugging, delegation, planning, code review, and configuration architecture - written from production experience and informed by the best open-source patterns.

The Problem

Claude Code is powerful. But without workflow discipline:

• Debugging = random guessing until something works
• Delegation = manual, inconsistent, context-wasting
• Planning = "just start coding" then rewrite 3 times
• Code review = skipped or superficial
• CLAUDE.md = unmanageable blob at 20+ rules

The Fix

5 skills that install in 30 seconds and immediately improve how you work:

Skill	What Changes
Debugging	4-phase cycle replaces random guessing. ACH method for complex bugs.
Delegation	Score-based routing to the right agent at the right model tier.
Plan & Execute	TDD plans that survive reality, with review gates between batches.
Code Review	Structured PR review with severity categorization and security focus.
Config Architect	Tiered CLAUDE.md that scales without wasting tokens.

Each skill includes reference files with templates, checklists, and examples.

Install

Full Bundle (Recommended)

git clone https://github.com/primeline-ai/primeline-skills ~/.claude-plugins/primeline-skills

Add to your Claude Code settings (~/.claude/settings.json):

{
  "pluginDirectories": [
    "~/.claude-plugins/primeline-skills"
  ]
}

Restart Claude Code. All 5 skills are now available automatically.

Individual Skills

If you only want one skill, copy its folder into your project's .claude/skills/ directory:

# Example: just the debugging skill
git clone https://github.com/primeline-ai/primeline-skills /tmp/primeline-skills
cp -r /tmp/primeline-skills/skills/debugging .claude/skills/

Verify Installation

Start Claude Code and ask it to debug something, plan a feature, or review code. It should announce "Using [skill-name]..." when the skill activates.

What's Inside Each Skill

Every skill follows the same structure:

skills/[name]/
  SKILL.md              # The skill itself (Quick Start + Full System)
  references/
    [template].md       # Ready-to-use templates
    [guide].md          # Deep-dive reference material

Skill	References
Debugging	Evidence collection template, anti-patterns guide
Delegation	5 prompt templates, model selection matrix
Plan & Execute	Plan template (full + minimal), review checklist
Code Review	Severity classification guide, security checklist
Config Architect	Copy-paste CLAUDE.md starter template, token measurement guide

The Ecosystem

These skills are part of a progression. Each tier works independently - no hard dependencies.

You're here          You want this            Install this
-----------          -------------            ------------
Raw Claude Code  ->  Session memory       ->  Starter System (free)
                 ->  Workflow skills      ->  + Skills Bundle (free) <- you are here
                 ->  Deep planning        ->  + UPF (free)
                 ->  Deep analysis        ->  + Quantum Lens (free)
                 ->  AI-powered system    ->  + Course (paid)

Component	What It Does	Link
Starter System	Session memory, handoffs, context awareness	GitHub
Skills Bundle	5 workflow skills (this repo)	You're reading it
UPF	Universal Planning Framework with 4-stage deep planning	GitHub
Quantum Lens	Multi-perspective analysis + solution engineering (7 cognitive lenses)	GitHub
Course	Kairn + Synapse: AI-powered memory and knowledge graphs	primeline.cc

How They Work Together

Skills reference each other with soft "if available" language:

• Debugging mentions /remember for logging known failures (Starter System)
• Plan & Execute mentions routing subtasks via the delegation skill (this bundle)
• All skills mention Kairn upgrade paths for AI-powered memory (Course)

If the referenced component isn't installed, nothing breaks. The tip is just skipped.

What the Course Adds

Each skill documents a "With Kairn" section explaining what becomes possible with AI-powered memory: