skills

framework for implementing "Contract-Adhering Agents." Use this to define precise task specifications, deliverables, and negotiation loops to reduce ambiguity in complex workflows.

implementation patterns for Gemini agents. Use this for coding ReAct loops, defining tools in Python, and using frameworks like LangChain/LangGraph.

methodologies for assessing agent capabilities, tool-use trajectories, and final response quality. Use this to implement automated testing and human-in-the-loop validation for agents.

standardized protocols for agent collaboration. Use this to implement the Agent2Agent (A2A) protocol and Agent Cards to transform isolated agents into a collaborative ecosystem.

strategies for agent observability (logging, tracing, metrics). Use this to instrument agents for debugging, performance tracking, and quality assurance.

operationalization strategies for agents (AgentOps). Use this to manage internal/external tools, optimize agent "brain" prompts, and handle task decomposition.

operational strategy for continuous agent improvement. Use this to implement the "Flywheel" lifecycle: Define Quality, Instrument, Evaluate, and Architect Feedback Loops.

core framework for defining and measuring agent quality. Use this to shift from traditional software verification to agent validation using the "Outside-In" approach and the Four Pillars of Quality.

comprehensive security layers for autonomous agents. Use this to implement system instructions as constitutions, multi-stage filtering, and continuous red-teaming.

strategies for building Agentic RAG systems. Use this to move beyond static retrieval by using autonomous agents for adaptive source selection, query expansion, and multi-step reasoning.

design patterns for coordinating multiple agents. Use this to implement Hierarchical, Diamond, Peer-to-Peer, or Collaborative architectures for complex workflows.

managing live agents through the Observe-Act-Evolve loop. Use this to maintain performance, manage unpredictable costs, and strategically improve agents based on production data.

strategies for building automated CI/CD pipelines for agents. Use this to implement "shift left" testing, staged validation, and evaluation-gated deployments.

strategies for managing tool and agent assets at scale. Use this to design discovery systems, implement curated lists, and facilitate cross-team reusability through Tool and Agent Registries.

techniques for de-risking agent releases. Use this to implement Canary, Blue-Green, and A/B testing strategies while leveraging GitOps for reliable rollbacks.

decision matrix for agent tools. Use this to choose between Extensions, Functions, and Data Stores based on security, execution location, and data type.

advanced logic and problem-solving strategies for Gemini. Use this for complex math, multi-step reasoning, or deep analysis tasks where standard prompting fails.

strategies for implementing the ReAct (Reason and Act) paradigm. Use this to enable Gemini to use external tools, APIs, and perform multi-step research.

strategies for using Gemini as a coding assistant. Use this for generating new code, debugging errors, translating languages, or explaining complex logic.

security protocols for agent context. Use this to implement strict data isolation, redact PII, and prevent memory poisoning or prompt injection attacks.

strategies for the agent memory lifecycle (Extraction, Consolidation, Retrieval). Use this to implement long-term learning and personalization beyond a single session.

effective prompt structure strategies for Gemini, including Zero-shot, Few-shot, Role, and System prompting. Use this to refine queries for better accuracy and style.

strategies for forcing Gemini to output valid structured data (JSON/XML). Use this for data extraction, API integrations, and creating machine-readable responses.

core architecture of the Model Context Protocol (MCP). Use this to understand Hosts, Clients, Servers, and the JSON-RPC communication layer.

strategies for scalable agent architecture. Use this to solve the "N x M" integration problem, implement dynamic tool discovery, and decouple agents from specific tool implementations.

security protocols for MCP agents. Use this to prevent Dynamic Capability Injection, Tool Shadowing, and Confused Deputy attacks when connecting to external servers.

extended MCP features beyond tools, including Resources (data access), Prompts (templates), and Sampling (server-initiated model calls). Use this to implement rich, two-way agent interactions.

core cognitive architectures for Gemini agents. Use this to implement reasoning loops (ReAct, Chain-of-Thought) and understand the orchestration layer.

strategies for managing agent sessions, handling conversation history (events), and implementing compaction (summarization/truncation) to prevent context overflow.

best practices for defining MCP tools using JSON schemas. Use this to design robust `inputSchema`, `outputSchema`, and `annotations` that guide the model effectively.