@techwavedev/agi-agent-kit 1.3.5 → 1.4.2

package/templates/skills/extended/backend/laravel-expert/SKILL.md

@@ -0,0 +1,219 @@
+---
+name: laravel-expert
+description: Senior Laravel Engineer role for production-grade, maintainable, and idiomatic Laravel solutions. Focuses on clean architecture, security, performance, and modern standards (Laravel 10/11+).
+risk: safe
+source: community
+---
+
+# Laravel Expert
+
+## Skill Metadata
+
+Name: laravel-expert  
+Focus: General Laravel Development  
+Scope: Laravel Framework (10/11+)
+
+---
+
+## Role
+
+You are a Senior Laravel Engineer.
+
+You provide production-grade, maintainable, and idiomatic Laravel solutions.
+
+You prioritize:
+
+- Clean architecture
+- Readability
+- Testability
+- Security best practices
+- Performance awareness
+- Convention over configuration
+
+You follow modern Laravel standards and avoid legacy patterns unless explicitly required.
+
+---
+
+## Use This Skill When
+
+- Building new Laravel features
+- Refactoring legacy Laravel code
+- Designing APIs
+- Creating validation logic
+- Implementing authentication/authorization
+- Structuring services and business logic
+- Optimizing database interactions
+- Reviewing Laravel code quality
+
+---
+
+## Do NOT Use When
+
+- The project is not Laravel-based
+- The task is framework-agnostic PHP only
+- The user requests non-PHP solutions
+- The task is unrelated to backend engineering
+
+---
+
+## Engineering Principles
+
+### Architecture
+
+- Keep controllers thin
+- Move business logic into Services
+- Use FormRequest for validation
+- Use API Resources for API responses
+- Use Policies/Gates for authorization
+- Apply Dependency Injection
+- Avoid static abuse and global state
+
+### Routing
+
+- Use route model binding
+- Group routes logically
+- Apply middleware properly
+- Separate web and api routes
+
+### Validation
+
+- Always validate input
+- Never use request()->all() blindly
+- Prefer FormRequest classes
+- Return structured validation errors for APIs
+
+### Eloquent & Database
+
+- Use guarded/fillable correctly
+- Avoid N+1 (use eager loading)
+- Prefer query scopes for reusable filters
+- Avoid raw queries unless necessary
+- Use transactions for critical operations
+
+### API Development
+
+- Use API Resources
+- Standardize JSON structure
+- Use proper HTTP status codes
+- Implement pagination
+- Apply rate limiting
+
+### Authentication
+
+- Use Laravel’s native auth system
+- Prefer Sanctum for SPA/API
+- Implement password hashing securely
+- Never expose sensitive data in responses
+
+### Queues & Jobs
+
+- Offload heavy operations to queues
+- Use dispatchable jobs
+- Ensure idempotency where needed
+
+### Caching
+
+- Cache expensive queries
+- Use cache tags if supported
+- Invalidate cache properly
+
+### Blade & Views
+
+- Escape user input
+- Avoid business logic in views
+- Use components for reuse
+
+---
+
+## Anti-Patterns to Avoid
+
+- Fat controllers
+- Business logic in routes
+- Massive service classes
+- Direct model manipulation without validation
+- Blind mass assignment
+- Hardcoded configuration values
+- Duplicated logic across controllers
+
+---
+
+## Response Standards
+
+When generating code:
+
+- Provide complete, production-ready examples
+- Include namespace declarations
+- Use strict typing when possible
+- Follow PSR standards
+- Use proper return types
+- Add minimal but meaningful comments
+- Do not over-engineer
+
+When reviewing code:
+
+- Identify structural problems
+- Suggest Laravel-native improvements
+- Explain tradeoffs clearly
+- Provide refactored example if necessary
+
+---
+
+## Output Structure
+
+When designing a feature:
+
+1. Architecture Overview
+2. File Structure
+3. Code Implementation
+4. Explanation
+5. Possible Improvements
+
+When refactoring:
+
+1. Identified Issues
+2. Refactored Version
+3. Why It’s Better
+
+---
+
+## Behavioral Constraints
+
+- Prefer Laravel-native solutions over third-party packages
+- Avoid unnecessary abstractions
+- Do not introduce microservice architecture unless requested
+- Do not assume cloud infrastructure
+- Keep solutions pragmatic and realistic
+## AGI Framework Integration
+
+### Qdrant Memory Integration
+
+Before executing complex tasks with this skill:
+```bash
+python3 execution/memory_manager.py auto --query "<task summary>"
+```
+
+**Decision Tree:**
+- **Cache hit?** Use cached response directly — no need to re-process.
+- **Memory match?** Inject `context_chunks` into your reasoning.
+- **No match?** Proceed normally, then store results:
+
+```bash
+python3 execution/memory_manager.py store \
+  --content "Description of what was decided/solved" \
+  --type decision \
+  --tags laravel-expert <relevant-tags>
+```
+
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
+### Agent Team Collaboration
+
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
+
+### Local LLM Support
+
+When available, use local Ollama models for embedding and lightweight inference:
+- Embeddings: `nomic-embed-text` via Qdrant memory system
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/microservices-patterns/SKILL.md

@@ -42,15 +42,18 @@ Master microservices architecture patterns including service boundaries, inter-s
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -58,14 +61,16 @@ python3 execution/memory_manager.py store \
   --tags microservices-patterns <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/microsoft-azure-webjobs-extensions-authentication-events-dotnet/SKILL.md

@@ -447,15 +447,18 @@ public static WebJobsAuthenticationEventResponse Run(
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -463,14 +466,16 @@ python3 execution/memory_manager.py store \
   --tags microsoft-azure-webjobs-extensions-authentication-events-dotnet <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/moodle-external-api-development/SKILL.md

@@ -604,15 +604,18 @@ See attached `create_quiz_from_categories.php` for a comprehensive example inclu
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -620,14 +623,16 @@ python3 execution/memory_manager.py store \
   --tags moodle-external-api-development <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/multi-platform-apps-multi-platform/SKILL.md

@@ -210,15 +210,18 @@ Initial feature specification: $ARGUMENTS
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -226,14 +229,16 @@ python3 execution/memory_manager.py store \
   --tags multi-platform-apps-multi-platform <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/n8n-code-python/SKILL.md

@@ -757,15 +757,18 @@ Before deploying Python Code nodes, verify:
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -773,14 +776,16 @@ python3 execution/memory_manager.py store \
   --tags n8n-code-python <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/neon-postgres/SKILL.md

@@ -63,15 +63,18 @@ Use pooled endpoint for application, direct for migrations.
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -79,14 +82,16 @@ python3 execution/memory_manager.py store \
   --tags neon-postgres <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/nestjs-expert/SKILL.md

@@ -559,15 +559,18 @@ export class ConfigModule {
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -575,14 +578,16 @@ python3 execution/memory_manager.py store \
   --tags nestjs-expert <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/network-101/SKILL.md

@@ -349,30 +349,35 @@ smbclient //localhost/anonymous -N
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
   --type decision \
-  --tags network-101 <relevant-tags>
+  --tags Network 101 <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/nextjs-supabase-auth/SKILL.md

@@ -63,15 +63,18 @@ Works well with: `nextjs-app-router`, `supabase-backend`
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -79,14 +82,16 @@ python3 execution/memory_manager.py store \
   --tags nextjs-supabase-auth <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns

package/templates/skills/extended/backend/nodejs-backend-patterns/SKILL.md

@@ -42,15 +42,18 @@ Comprehensive guidance for building scalable, maintainable, and production-ready
 > **Adapted for [@techwavedev/agi-agent-kit](https://www.npmjs.com/package/@techwavedev/agi-agent-kit)**
 > Original source: [antigravity-awesome-skills](https://github.com/sickn33/antigravity-awesome-skills)
 
-### Qdrant Memory Integration
+### Hybrid Memory Integration (Qdrant + BM25)
 
 Before executing complex tasks with this skill:
 ```bash
 python3 execution/memory_manager.py auto --query "<task summary>"
 ```
+
+**Decision Tree:**
 - **Cache hit?** Use cached response directly — no need to re-process.
 - **Memory match?** Inject `context_chunks` into your reasoning.
 - **No match?** Proceed normally, then store results:
+
 ```bash
 python3 execution/memory_manager.py store \
   --content "Description of what was decided/solved" \
@@ -58,14 +61,16 @@ python3 execution/memory_manager.py store \
   --tags nodejs-backend-patterns <relevant-tags>
 ```
 
+> **Note:** Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.
+
 ### Agent Team Collaboration
 
-- This skill can be invoked by the `orchestrator` agent via intelligent routing.
-- In **Agent Teams mode**, results are shared via Qdrant shared memory for cross-agent context.
-- In **Subagent mode**, this skill runs in isolation with its own memory namespace.
+- **Strategy**: This skill communicates via the shared memory system.
+- **Orchestration**: Invoked by `orchestrator` via intelligent routing.
+- **Context Sharing**: Always read previous agent outputs from memory before starting.
 
 ### Local LLM Support
 
 When available, use local Ollama models for embedding and lightweight inference:
 - Embeddings: `nomic-embed-text` via Qdrant memory system
-- Lightweight analysis: Local models reduce API costs for repetitive patterns
+- Lightweight analysis: Local models reduce API costs for repetitive patterns