a5c-ai

Project context loading, isolation, and persistent state management across CCPM sessions.

CrewAI multi-agent orchestration setup for collaborative AI systems

Generate personalized verified news briefs with 7-day freshness and 95%+ source accuracy

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies.

Entity and fact extraction for user profiling and personalization

Transform PRD into a technical implementation plan with architecture decisions and dependency mapping.

Log all errors with full context, detect patterns, and suggest approach mutations to avoid repeated failures.

Parse and normalize features from text descriptions, images, and screenshots into structured requirements.

Few-shot example generation and optimization for improved LLM performance

Capture and persist research findings, discoveries, and decision rationale to findings.md.

Use when implementation is complete, all tests pass, and you need to decide how to integrate the work.

Bidirectional synchronization of epics and tasks with GitHub issues, labels, and relationships.

Haystack NLP pipeline configuration for document processing and QA

Urgent issue classification, root cause analysis, and fast-path routing for production hotfixes

Hugging Face transformer model fine-tuning and inference for intent classification

Initialize and manage Kanban board state for feature workflow tracking.

Build structured knowledge frameworks from scattered vault notes with source attribution

Capture, validate, query, and sync architectural patterns and design decisions in the knowledge graph

LangChain chain composition including SequentialChain, RouterChain, and LCEL patterns

LangChain memory integration including ConversationBufferMemory, ConversationSummaryMemory, and vector-based memory

LangChain ReAct agent implementation with tool binding for reasoning and action loops

LangChain retriever implementation with various retrieval strategies for RAG applications

LangChain tool creation and integration utilities for agent systems

LangFuse LLM observability integration for tracing, analytics, and cost tracking

LangGraph checkpoint and persistence configuration for stateful workflow management

Human-in-the-loop integration for LangGraph workflows with approval and intervention points

Conditional edge routing and state-based transitions for LangGraph workflows

Subgraph composition and modular workflow design for LangGraph

LlamaIndex agent and query engine setup for RAG-powered agents

LLM-based zero-shot and few-shot classification for flexible intent detection

Technical debt management including branch cleanup, doc verification, TODO scanning, and dependency auditing

Process meeting recordings and transcripts into decisions, action items, and team dynamics

Conversation summarization for memory compression and context management

Milvus distributed vector database configuration for large-scale RAG applications

NVIDIA NeMo Guardrails configuration for conversational safety and control

Personalize COG Second Brain workflow through role pack selection and vault initialization

OpenTelemetry instrumentation for LLM applications with distributed tracing

Coordinate multiple specialized agents across parallel work streams with quality gates and conflict resolution.

Continuous monitoring using Deacon/Witness patterns for agent health checks, stuck detection, and automated recovery.

Arize Phoenix observability platform setup for LLM debugging and evaluation

PII detection and redaction utilities for privacy-compliant conversational AI

Pinecone vector database setup, configuration, and operations for RAG applications

Create a structured task_plan.md with phases, goals, and checkbox tracking for persistent planning.

Create Product Requirements Documents through structured discovery using Jobs-to-be-Done framework.

Create comprehensive product briefs from market, domain, and technical research.

Status dashboards, standup reports, and blocked task management with GitHub synchronization.

Generate comprehensive project documentation with source tree analysis and Mermaid diagrams.

Token-efficient prompt compression techniques for cost optimization

Prompt injection detection and prevention for secure LLM applications

Structured prompt template creation with variables, formatting, and version control

Generate and execute API and E2E tests with quality gate assessment.

Qdrant vector database with filtering, payloads, and quantization support

Lean solo-developer workflow for rapid spec-to-implementation with minimal overhead.

Document chunking with multiple strategies including semantic, recursive, and fixed-size chunking

Batch embedding generation with caching, rate limiting, and multiple provider support

Query expansion, HyDE, and multi-query generation for improved retrieval

Cross-encoder reranking and MMR diversity filtering for improved retrieval quality

Rasa NLU pipeline configuration and training for intent and entity extraction

Redis backend for conversation state persistence and caching

Use when completing tasks, implementing major features, or before merging to verify work meets requirements.