Laraskills
LaraSkillsContext engine online

Laravel engineeringcontext forAI coding agents

LaraSkills equips OpenCode, Claude Code, Cursor, Codex, Copilot, and other coding agents with Laravel-specific skills, rules, retrieval workflows, and MCP support so they can work with deeper context and fewer guesses.

Install LaraSkillsView GitHub
202 tests passed2,321 knowledge unitsMCP ready
laraskills.context
MCP

Retrieval session

Optimize N+1 query

Ready
bash00:02.4
$npm install --save-dev laraskills
$laraskills retrieve "Optimize N+1 query"
Found Laravel-specific context
Loaded rules and examples
Context ready for your AI agent

Skills

4

Rules

12

Examples

8

How it works

Your agent writes the code.Laraskills supplies the Laravel judgment.

Laraskills turns a plain-language task into a compact, relevant context packet before generation starts, so your coding agent works from framework-specific guidance instead of broad memory.

01Request
Developer intent
task.md
“Build a secure profile update flow with authorization and validation.”
Plain-language task
Any coding agent
02Retrieve

Context packet

Assembled for this task, not every task.

Ready
  1. 01

    Route the task

    Identify the Laravel domains and decisions involved.

  2. 02

    Load the right guidance

    Select focused skills, rules, examples, and checklists.

  3. 03

    Resolve dependencies

    Include prerequisites and related framework constraints.

  4. 04

    Return a bounded packet

    Give the agent useful context without flooding its window.

SkillsRulesExamplesDependencies
03Build
Agent output
context loaded
ProfileController.php
public function update(
  UpdateProfileRequest $request
): RedirectResponse
  • Laravel conventions are explicit
  • Security decisions are considered
  • Less prompting and fewer guesses

Input

One clear engineering task

Context

The smallest useful Laravel guidance

Result

Code grounded in framework decisions

Scale

LaraSkills at a glance

The LaraSkills knowledge system today

0Knowledge units
Core retrieval foundation

Curated Laravel knowledge chunks that ground AI coding agents with precise, framework-aware context.

Knowledge Core
0Laravel skills

Implementation skills across key Laravel engineering tasks and patterns.

0Always-follow rules

Guardrails that improve consistency and reduce unsafe output.

Ecosystem Coverage
0Agent definitions

Structured agent behaviors for coding and retrieval workflows.

0Engineering domains

Coverage across architecture, validation, APIs, queues, testing, and more.

Integration Layer
0Read-only MCP tools

Safe protocol tools that expose LaraSkills intelligence to coding agents.

0Supported coding harnesses

Compatible with multiple agent environments and developer workflows.

Before & after

Generic coding agent vs Laraskills

Why Laravel developers get better results with Laraskills

Generic coding agent

  • ✕Guesses Laravel conventions
  • ✕Relies on broad framework memory
  • ✕Misses architecture decisions
  • ✕Needs long manual prompts

Agent with Laraskills

  • ✓Loads Laravel-specific guidance
  • ✓Uses focused retrieval
  • ✓Follows skills, rules, and checklists
  • ✓Works with MCP-compatible tools
Structure

Three-layer architecture

How Laraskills structures its engineering knowledge

1

Operating layer

  • Skills, rules, agents
  • Command references
  • Harness configs
2

Intelligence layer

  • Domains, knowledge units
  • Indexes, routing maps
  • Dependency graph
3

Retrieval layer

  • CLI retrieval, search
  • Inspect, graph context
  • MCP tools
Terminal

CLI at your fingertips

Real commands to install, retrieve, search, inspect, and diagnose — all from your terminal

Install Laraskills
$ npm install --save-dev laraskills
Retrieve context
$ npx laraskills retrieve "Optimize an N+1 query" --mode compact
Search knowledge
$ npx laraskills search "Policies versus Gates"
Inspect knowledge unit
$ npx laraskills get security-identity-engineering/authentication/sanctum-spa-authentication
Validate knowledge graph
$ npx laraskills validate
System diagnostics
$ npx laraskills doctor
Protocol

MCP tools for coding agents

Laraskills exposes read-only, deterministic, local-retrieval MCP tools for compatible coding agents

retrieve_context_bundleRead-only

Return the smallest useful Laraskills context bundle for a Laravel engineering task.

search_eccRead-only

Search the Laraskills knowledge unit catalog with ranked results.

get_knowledge_unitRead-only

Inspect a single canonical knowledge unit by ID with bounded content.

get_graph_contextRead-only

Return prerequisites and related topics for a knowledge unit.

validate_eccRead-only

Validate the structural integrity of the Laraskills intelligence layer.

Ecosystem

Works with your Laravel AI stack

Use Laraskills with the coding agents, editors, and assistants already in your workflow.

OpenCode
Claude Code
Cursor
Gemini CLI
Codex CLI
GitHub Copilot
VS Code
Zed
Trae
Qwen
CodeBuddy
Kiro
Agent CLIs

Run Laraskills-guided Laravel workflows directly from your terminal.

OpenCode
Gemini CLI
Codex CLI
Qwen
AI IDEs and editors

Bring Laravel-specific context into the editors where your code lives.

Cursor
VS Code
Zed
Trae
AI assistants

Give assistant-style tools clearer Laravel rules, retrieval, and implementation context.

Claude Code
GitHub Copilot
CodeBuddy
Kiro

Tool names and logos are trademarks of their respective owners.

Evidence

Research & evidence

In an attribution study, Laraskills was tested against baseline coding agents across multiple Laravel implementation scenarios. The results are measured and reproducible.

9Isolated OpenCode runs
3Laravel implementation scenarios
100%Test & Pint pass rate
2/3Required-context mode won

These results are from a controlled study and do not claim Laraskills always outperforms baseline agents. Performance depends on task complexity, agent choice, and configuration.

Read the research

Ready to ship better Laravel code?

Install Laraskills and give your coding agent Laravel-specific context in seconds.

Terminal
$ npm install --save-dev laraskills
Install LaraskillsView GitHub