Code Spice
Code quality spice for Line Cook — decision frameworks, antipattern detection, refactoring patterns, tradeoff analysis, YAGNI guidance, dead code pruning, plan auditing, and a code quality critic agent.
Quick Start
After installing Line Cook:
/plugin marketplace add smileynet/line-cook
/plugin install code@line-cook
/line:mise # Plan a projectUpdate: /plugin update code
What It Looks Like
Here's a scoping session with code-spice active. The skills inject code-quality-specific questions and frameworks automatically:
You: I need to add a payment processing module.
Claude: Let's think through the quality dimensions:
**What error handling strategy fits this module?** ← error-handling-patterns
Payments are critical — errors must be explicit.
┌──────────────────┬─────────────────────────────────┐
│ Exceptions │ Good for truly exceptional cases │
│ Result types │ Best for expected failures │
│ Error codes │ Low-level, C-style APIs │
│ Nullable │ Avoid for critical paths │
└──────────────────┴─────────────────────────────────┘
For payment processing, Result types make failures
explicit and force callers to handle them.
You: Result types. What about the API design?
Claude: Good. Now let's evaluate the tradeoffs:
**How flexible should the payment API be?** ← software-tradeoffs
┌─────────────────────┬──────────────────────────────┐
│ Simple & rigid │ Fast to build, hard to extend │
│ Flexible & complex │ Handles edge cases, more code │
│ Plugin-based │ Extensible but over-engineered │
└─────────────────────┴──────────────────────────────┘
YAGNI says start simple. But payment providers change —
a thin adapter layer gives flexibility without complexity.Each question comes from a different skill, but they flow as a natural conversation.
When to Use
Use code-spice when:
- Planning a software project with
/line:mise - Making design decisions about error handling, API shape, or abstractions
- Implementing code during
/line:cook - Running
/line:plan-auditto check plan quality - Running
/line:servefor enhanced code review with quality criteria - Running
/code:tradeoff,/code:smell,/code:review-prep, or/code:prune
Skip it when:
- Building a game (use game-spice instead)
- Working on infrastructure or ops tasks
Commands
| Command | Description | When to Use | |---------|-------------|-------------| | /code:tradeoff | Systematic tradeoff analysis for design decisions | Evaluating competing approaches | | /code:smell | Structured code smell detection on recent changes | Catching antipatterns before review | | /code:review-prep | Context-aware self-review checklist | Preparing for code review | | /code:prune | Interactive codebase pruning analysis | Dead code, unused deps, safe removal |
Agents
| Agent | Description | |-------|-------------| | code-quality-critic | Reviews code quality against six pillars during /line:serve. |
What's Inside
| Skill | Activates During | Key Content | |-------|-------------------|------------| | code-quality-foundations | /brainstorm, /scope | 4 goals table, 6 pillars table, abstraction/investment decision tables | | code-readability | /scope, /cook | Comment decision table, readability checklist, documentation hierarchy | | code-naming | /scope, /cook | Naming decision table, three questions test, antipattern tables | | refactoring-patterns | /scope, /cook | 13-pattern catalog, smell-to-pattern mapping, invariant hierarchy | | error-handling-patterns | /scope, /cook | Strategy decision table, recoverability framework, signaling comparison | | code-antipatterns | /scope, /plan-audit | Top-10 severity table, antipattern-or-tradeoff table, context evaluation | | code-review | /serve, /cook | Review goals, Triple-R pattern, severity labels, TWA template, 8 antipatterns | | code-testing-quality | /cook | Quality checklist, test doubles table, antipattern catalog, coverage guidance | | software-tradeoffs | /brainstorm, /scope | Tradeoff matrix, "which tradeoff?" table, "chose wrong?" diagnostics | | code-plan-audit | /plan-audit | 10-point scorecard, actionability tests, plan-level antipatterns | | code-yagni | /brainstorm, /scope | Build decision table, Kohavi stats, speculative generality signals | | code-scope-boundaries | /brainstorm, /scope | Scope health checklist, feature belonging test, split-vs-keep tree | | code-pruning | /cook, /architecture-audit | SCARF pattern, tool table, safe removal process, bloat metrics |
13 skills (~1,200 lines), 4 commands, 1 agent. Each skill is under 150 lines — focused on decision tables and checklists that change Claude's behavior, not prose Claude already knows.
How It Works
Skills load automatically when Line Cook commands detect code project context.
/brainstorm /scope /plan-audit
┌────────────────────┐ ┌──────────────────────────┐ ┌──────────────────┐
│ code-quality- │ │ code-readability │ │ code-antipatterns │
│ foundations │ │ code-naming │ │ code-plan-audit │
│ software-tradeoffs │───>│ refactoring-patterns │───>│ │
│ code-yagni │ │ error-handling-patterns │ └──────────────────┘
│ code-scope- │ │ code-antipatterns │
│ boundaries │ │ software-tradeoffs │
│ │ │ code-yagni │
│ │ │ code-scope-boundaries │
└────────────────────┘ └──────────────────────────┘
/line:cook /line:serve
┌────────────────────┐ ┌──────────────────────────┐
│ code-readability │ │ code-review │
│ code-naming │ │ code-quality-critic │
│ refactoring- │ │ (agent) │
│ patterns │ └──────────────────────────┘
│ error-handling- │
│ patterns │
│ code-review │
│ code-testing- │
│ quality │
│ code-pruning │
└────────────────────┘Source Material
Knowledge synthesized from four code quality books:
| Book | Key Contributions | |------|-------------------| | Good Code, Bad Code (Tom Long, 2021) | Quality pillars, error handling, antipatterns | | Five Lines of Code (Christian Clausen, 2021) | Refactoring catalog, rule-based triggers | | Software Mistakes and Tradeoffs (Lelek & Skeet, 2022) | Tradeoff frameworks, API design | | Looks Good to Me (Braganza, 2025) | Code review process, AI-augmented review |
FAQ
Do I need to configure anything? No. Install the plugin and plan a project. Skills activate automatically.
Does code-spice work without Line Cook? The skills are designed for Line Cook's /line:mise workflow. Without Line Cook, Claude won't have the planning commands that trigger these skills.
What languages does it cover? The guidance is language-agnostic. Principles apply to any language.
Can I use code-spice with game-spice? Yes. They cover different domains and activate on different keywords.
Learn More
- Changelog — Version history
- Skill Content Guide — What to put in a skill (and what to leave out)
Local Development
/plugin marketplace add /home/sam/code/code-spiceRequirements
- Claude Code >= 1.0.33
- Line Cook plugin
License
MIT
