AI Agents
Thông số và tài liệu đọc được bởi máy, thiết kế riêng cho trình phân tích cú pháp và bot AI.
72 articles · Page 1 of 6
Agent Authentication Documentation Spec
Document authentication for autonomous agents: OAuth flows, API keys, scopes, error states, and consent UX patterns AI agents need to operate safely.
Agent Circuit Breaker Specification
Specification for circuit breakers protecting AI agent calls to LLM providers and tools, including state transitions, threshold tuning, fallback strategies, and observability hooks.
Agent Citation Attribution Specification: Verifiable Source Tracking for Autonomous AI Agents
Specification defining HTTP headers, provenance manifests, and chain-of-citation markup so autonomous AI agents produce verifiable citations to source content.
Agent Context Window Budgeting Specification
Agent context window budgeting spec: token allocation buckets, summarization triggers, eviction policies, prompt caching pairing, and worked examples.
Agent Conversation Summarization: Triggers, Schema, and Retention
Specification for compressing agent conversation history into running summaries: triggers, summary schema, retention rules, and recovery patterns for long-running chats.
Agent Cost Disclosure Specification
Specification for declaring per-call and per-token costs in AI agent tool manifests so cost-aware planners can budget runs and avoid Denial-of-Wallet failures.
Agent End-to-End Testing Specification
E2E testing spec for AI agents: scenario suites, golden traces, LLM-as-judge evaluators, deterministic replay, and CI gating thresholds.
Agent Error Handling Documentation Specification: Designing Errors Agents Can Self-Repair From
Spec for documenting error states, validation messages, and self-repair hints so AI agents recover automatically when calling your tools and APIs.
Agent Error Recovery Patterns Specification
Specification for agent error recovery — retry strategies, idempotency keys, Saga compensation, poison-message handling, and runbook-friendly error codes.
Agent Evaluation Harness Documentation: How to Spec an Eval Suite for AI Agents
Specification for documenting an AI agent evaluation harness — eval suites, scorers, datasets, and trajectory grading that humans and docs agents can both consume.
Agent-Friendly Documentation: How to Design Docs for Autonomous Agents
Specification for agent-friendly documentation: structure, affordances, tool contracts, deterministic anchors, and anti-patterns that break autonomous agents.
Agent Graceful Degradation Specification
Specification for graceful degradation when AI agent dependencies fail: model fallback chains, tool-skip policies, cached-response serving, and user-facing failure messaging.