AI Agents · Page 2
Thông số và tài liệu đọc được bởi máy, thiết kế riêng cho trình phân tích cú pháp và bot AI.
72 articles · Page 2 of 6
Agent Hallucination Detection Spec
A specification for detecting hallucinations inside AI-agent runs by combining tool-call verification, groundedness scoring, and self-consistency checks before the agent emits a final answer.
Agent Handoff Protocol Documentation Spec for Multi-Agent AI Systems
Specification for documenting agent handoff protocols in multi-agent AI systems—trigger conditions, context payload, idempotency, and recovery.
Agent Health Check Specification
Specification for liveness, readiness, and startup probes in production AI agents, including LLM-provider ping patterns, dependency probing, and degraded-mode signaling.
Agent Idempotency Documentation Specification
Specification for documenting idempotency keys, retry semantics, and safe-replay guarantees so AI agents recover from failures without duplicates.
Agent Knowledge Base Integration: RAG, MCP, and Direct API Patterns
Spec for connecting AI agents to internal knowledge bases via RAG vector stores, MCP servers, or direct retrieval APIs with provenance and ACL stamping.
Agent Knowledge Base Specification: Structure, Refresh, and Versioning
Production specification for AI agent knowledge bases: document model, chunking strategies, metadata enrichment, refresh cadence, version pinning, and rollback.
Knowledge Cutoff Disclosure Specification for AI Agent Documentation
A normative specification for disclosing AI agent knowledge cutoff dates in documentation, with required fields, examples, and validation rules.
Agent Output License Disclosure Specification
Spec for disclosing agent output licensing—training-data provenance, output-rights statement, and downstream attribution requirements.
Agent Long-Running Job Documentation Specification
Specification for documenting long-running agent jobs: async kickoff, status polling, SSE progress, cancellation, and timeout SLAs that AI agent tools must publish.
Agent MCP Server Discovery Specification
How AI agents discover, register, and connect to Model Context Protocol (MCP) servers — manifests, capability handshake, auth, and trust verification.
Agent Memory Architecture for Long-Running Documentation Agents
Memory architecture spec for long-running documentation agents: working, episodic, semantic, and procedural layers tuned for citation-stable publishing.
Agent Memory Pattern Specification: Short-Term, Long-Term, and Episodic
Specification for AI agent memory: working, episodic, semantic, and procedural tiers with consolidation, eviction, and PII handling.