Shipkit Core Laravel Package

correct

Find the mistakes agents keep repeating in this repo and make each one impossible: architecture first, then a failing check, then a test, docs last. Prove each check fails on a real past mistake. Run after the user corrects an agent for the same thing twice.

claude-code

diagnosing-bugs

Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.

claude-code

domain-modeling

Build and sharpen a bounded context's domain model. Use when discussing codebase terminology, defining a new context, writing a context's vocabulary or invariants in .ai/rules/contexts, or recording an ADR.

claude-code

grilling

Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.

claude-code

grill-with-docs

A relentless interview to sharpen a plan or design, which also updates the context rule files and ADRs as we go.

claude-code

handoff

Compact the current conversation into a handoff document for another agent to pick up.

claude-code

implement

Implement a piece of work based on a spec or set of GitHub issue tickets, test-first, finishing with composer ci:check and a review.

claude-code

new-context

Create a new bounded context in app/Domain. Use when a feature introduces its own business vocabulary or invariants, such as billing, catalog, or teams, rather than extending an existing context.

claude-code

retro

Suggest improvements to the coding agent's environment (navigation, automated checks, rules, tooling) after a session, most severe first.

claude-code

review-diff

Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes: Standards (does the code follow this repo's .ai rules?) and Spec (does it match the originating issue?). Runs both reviews in parallel sub-agents. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".

claude-code

to-spec

Turn the current conversation into a spec and publish it as a GitHub issue: no interview, just synthesis of what you've already discussed.

claude-code

to-tickets

Break a plan, spec, or the current conversation into tracer-bullet tickets, each declaring its blocking edges, published as GitHub issues.

claude-code

writing-for-agents

Writing documents for agents. Use when creating or editing skills, .ai/guidelines, or .ai/rules files.

claude-code

Writing For Agents Skill Mechanics

The skill-specific branch of writing-for-agents: what changes when the document is a skill (frontmatter, the invocation choice, and router skills). Everything else about writing it is the universal reference in SKILL.md.

claude-code

Domain Modeling Vocabulary Format

Each bounded context's model lives in .ai/rules/contexts/{context}.md, scaffolded by php artisan make:context. Keep its three sections in this order.

claude-code

Domain Modeling Adr Format

ADRs live in docs/adr/ and use sequential numbering: 0001-slug.md, 0002-slug.md, etc.

claude-code

Review Diff Checklist

The reviewer's standards beyond what composer ci:check already enforces. Rows of .ai/rules/enforcement.md that name a check are covered by that check; spend review attention on the rest.

claude-code