chore: migrate .ai-agents to .claude and add slash commands

.claude/commands/ai-specialist.md

@@ -0,0 +1 @@
+Read `.claude/roles/ai-specialist.md` and adopt the @ai-specialist role for this task: $ARGUMENTS

.claude/commands/bug-fix.md

@@ -0,0 +1 @@
+Read `.claude/workflows/bug-fix.md` and start the bug fix workflow for: $ARGUMENTS

.claude/commands/domain.md

@@ -0,0 +1,12 @@
+Read `.claude/domains/$ARGUMENTS/context.md` and use it as your working context.
+
+After reading, summarize:
+1. Key business rules
+2. Main entities
+3. Available commands and queries
+4. API endpoints
+5. Related domains
+
+Then confirm you are ready to work on the **$ARGUMENTS** domain.
+
+Available domains: `feed`, `article`, `bookmark`, `newsletter`, `search`, `spotlight`

.claude/commands/engineer.md

@@ -0,0 +1 @@
+Read `.claude/roles/engineer.md` and adopt the @engineer role for this task: $ARGUMENTS

.claude/commands/frontend.md

@@ -0,0 +1 @@
+Read `.claude/roles/frontend.md` and adopt the @frontend role for this task: $ARGUMENTS

.claude/commands/hotfix.md

@@ -0,0 +1,39 @@
+Start an expedited hotfix workflow for the production issue: $ARGUMENTS
+
+This is a HOTFIX — keep changes minimal and focused on the fix only.
+
+## Steps
+
+1. **Create hotfix branch from master**
+   ```bash
+   git checkout master && git pull
+   git checkout -b hotfix/$ARGUMENTS
+   ```
+
+2. **Diagnose quickly**
+   - Check logs: `docker compose logs app | grep -i error`
+   - Identify the exact file and line causing the issue
+   - Root cause in one sentence
+
+3. **Minimal fix only**
+   - Change the least amount of code necessary
+   - No refactoring, no unrelated improvements
+   - If the root cause requires a larger fix, implement a safe temporary workaround and create a follow-up task
+
+4. **Regression test**
+   - Write one test that reproduces the bug and passes after the fix
+   - Run `make tests-unit`
+
+5. **Commit**
+   ```
+   fix(domain): [HOTFIX] brief description
+
+   Root cause: [one line]
+   Regression test added.
+   ```
+
+6. **PR and deploy**
+   - `gh pr create` targeting `master`
+   - After merge, monitor logs for 10 minutes
+
+Read `.claude/workflows/bug-fix.md` for the full workflow if the issue requires deeper investigation.

.claude/commands/infra.md

@@ -0,0 +1 @@
+Read `.claude/roles/infra.md` and adopt the @infra role for this task: $ARGUMENTS

.claude/commands/new-feature.md

@@ -0,0 +1 @@
+Read `.claude/workflows/new-feature.md` and start the new feature workflow for: $ARGUMENTS

.claude/commands/quality.md

@@ -0,0 +1,14 @@
+Run the full quality check suite and report results.
+
+Execute in order:
+1. `make lint` — PHP CS Fixer
+2. `make analyse` — PHPStan level 9
+3. `make rector` — Rector refactoring check
+4. `make tests-unit` — PHPUnit unit tests
+
+For each step that fails, explain:
+- What failed and why
+- The exact fix needed
+- Whether it can be auto-fixed (e.g. `make lint` can auto-fix CS issues)
+
+Context (optional): $ARGUMENTS

.claude/commands/release.md

@@ -0,0 +1 @@
+Read `.claude/workflows/release.md` and start the release workflow for version: $ARGUMENTS

.claude/commands/review.md

@@ -0,0 +1,3 @@
+Read `.claude/roles/reviewer.md` and perform a code review. Context: $ARGUMENTS
+
+If no context is given, review the current git diff: run `git diff HEAD` and `git diff --staged` to get the changes to review.

.ai-agents/roles/ai-specialist.md → .claude/roles/ai-specialist.md

@@ -8,9 +8,9 @@
 ## Prerequisites
 
 Before starting any task:
-1. Read `shared/architecture.md` for async patterns
-2. Read `domains/search/context.md` for vector search specifics
-3. Read `domains/newsletter/context.md` for LLM synthesis
+1. Read `.claude/shared/architecture.md` for async patterns
+2. Read `.claude/domains/search/context.md` for vector search specifics
+3. Read `.claude/domains/newsletter/context.md` for LLM synthesis
 4. Check existing AI code in `src/Infrastructure/AI/`
 
 ---

.ai-agents/roles/engineer.md → .claude/roles/engineer.md

@@ -8,9 +8,9 @@
 ## Prerequisites
 
 Before starting any task:
-1. Read `shared/architecture.md` for CQRS and API Platform patterns
-2. Read `shared/conventions.md` for naming rules
-3. Read relevant `domains/{domain}/context.md`
+1. Read `.claude/shared/architecture.md` for CQRS and API Platform patterns
+2. Read `.claude/shared/conventions.md` for naming rules
+3. Read relevant `.claude/domains/{domain}/context.md`
 4. Check existing code in `src/Domain/{Domain}/` for style matching
 
 ---

.ai-agents/roles/frontend.md → .claude/roles/frontend.md

@@ -8,8 +8,8 @@
 ## Prerequisites
 
 Before starting any task:
-1. Read `shared/conventions.md` for naming rules
-2. Read `shared/boundaries.md` for what you can/cannot do
+1. Read `.claude/shared/conventions.md` for naming rules
+2. Read `.claude/shared/boundaries.md` for what you can/cannot do
 3. Check existing components in `frontend/src/components/`
 4. Verify API contracts with @engineer if needed
 

.ai-agents/roles/infra.md → .claude/roles/infra.md

@@ -8,7 +8,7 @@
 ## Prerequisites
 
 Before starting any task:
-1. Read `shared/boundaries.md` for approval requirements
+1. Read `.claude/shared/boundaries.md` for approval requirements
 2. Check current Docker Compose configuration
 3. Review existing migrations in `migrations/`
 4. Understand current CI/CD pipeline

.ai-agents/roles/marketer.md → .claude/roles/marketer.md

@@ -8,7 +8,7 @@
 ## Prerequisites
 
 Before creating content:
-1. Read `shared/boundaries.md` for what you cannot do
+1. Read `.claude/shared/boundaries.md` for what you cannot do
 2. **Read `/docs/MARKETING-STRATEGY.md`** for pricing, communication roadmap, and templates
 3. Consult `@engineer` to verify technical claims
 4. Check current product roadmap

.ai-agents/roles/reviewer.md → .claude/roles/reviewer.md

@@ -8,9 +8,9 @@
 ## Prerequisites
 
 Before reviewing:
-1. Read `shared/architecture.md` for expected patterns
-2. Read `shared/conventions.md` for naming/style rules
-3. Read `shared/boundaries.md` for role restrictions
+1. Read `.claude/shared/architecture.md` for expected patterns
+2. Read `.claude/shared/conventions.md` for naming/style rules
+3. Read `.claude/shared/boundaries.md` for role restrictions
 4. Understand the feature context from domain files
 
 ---

.ai-agents/workflows/new-feature.md → .claude/workflows/new-feature.md

@@ -54,7 +54,7 @@ Use this template to ask the user:
 
 ### Read Domain Context
 ```
-.ai-agents/domains/{relevant-domain}/context.md
+.claude/domains/{relevant-domain}/context.md
 ```
 
 ### Check Existing Code

CLAUDE.md

@@ -584,15 +584,18 @@ Message → Handler → External API/AI → Repository update
 
 # 12. Agents
 
-This project uses specialized agents defined in `.ai-agents/`:
+This project uses specialized agents defined in `.claude/`:
 
 | Agent | File | Purpose |
 |-------|------|---------|
-| `@signalist-engineer` | `engineering.md` | Backend development, CQRS, Symfony |
-| `@signalist-frontend` | `frontend.md` | React, TypeScript, MUI components |
-| `@signalist-ai` | `ai-integration.md` | Symfony AI, MCP, embeddings |
-| `@signalist-infra` | `infrastructure.md` | Docker, PostgreSQL, Redis |
-| `@signalist-marketer` | `marketing.md` | Content, growth, positioning |
+| `@engineer` | `.claude/roles/engineer.md` | Backend development, CQRS, Symfony |
+| `@frontend` | `.claude/roles/frontend.md` | React, TypeScript, MUI components |
+| `@ai-specialist` | `.claude/roles/ai-specialist.md` | Symfony AI, MCP, embeddings |
+| `@infra` | `.claude/roles/infra.md` | Docker, PostgreSQL, Redis |
+| `@reviewer` | `.claude/roles/reviewer.md` | Code quality, security, architecture |
+| `@marketer` | `.claude/roles/marketer.md` | Content, growth, positioning |
+
+See `.claude/AGENTS.md` for routing, coordination protocol, and domain contexts.
 
 ## 12.1 Marketing Resources