AI Agents · Page 4
Thông số và tài liệu đọc được bởi máy, thiết kế riêng cho trình phân tích cú pháp và bot AI.
72 articles · Page 4 of 6
Agent Retry Strategy Specification
Retry-strategy specification for AI agents covering retry classes, exponential backoff with jitter, idempotency keys, Retry-After honoring, and per-tenant retry budgets.
Agent Sandbox Documentation Specification: Documenting Execution Environments for Autonomous AI Agents
Documentation specification for describing AI agent sandbox environments — isolation tier, filesystem, ports, snapshots, resource limits, network policy, and tool access — so downstream agents and operators can use them safely.
Agent Secret Management Specification
Specification for agent secret management — vault storage, dynamic / short-lived credentials, rotation, tool-scoped access, and exposure prevention.
Agent Self-Correction Loop: Critique, Revise, and Converge
Spec for agent self-correction loops: critique step, revision policy, max iterations cap, convergence test, and same-model vs separate-critic tradeoffs.
Agent Skill Manifest Specification: Publishing SKILL.md for AI Agent Discovery
Agent Skill Manifest specification: how to author and publish SKILL.md so Claude, ChatGPT, Codex, Gemini, and Copilot agents discover and reuse your docs.
Agent Startup and Shutdown Specification
Specification for agent startup, warmup, draining, and graceful shutdown lifecycle hooks ensuring zero-message-loss during deploys and predictable cold-start behavior.
Agent State Management Patterns Specification
Specification for agent state management — short/long/durable state, storage backends, checkpointing, and crash-recovery semantics.
Agent Streaming Output Documentation Spec: Events, Errors, Partial State
Vendor-neutral specification for documenting AI agent streaming output: event taxonomy, partial state, error frames, cancellation, and resume semantics.
Agent Tool Latency Disclosure Specification: Documenting Response Time SLAs for AI Agent Tools
Agent tool latency disclosure spec: a documentation standard for declaring p50/p95/p99 response time SLAs of tools called by AI agents.
Agent Tool Manifest QA Checklist: Validating SKILL.md and Tool Schemas Before Agent Discovery
QA checklist for agent tool manifests: validate SKILL.md, JSON schemas, descriptions, and examples so agents discover and call tools correctly.
Agent Tool Naming Conventions Specification for LLM Routing Reliability
Specification for naming AI agent tools to maximize LLM routing reliability: verb-noun, namespaces, length, anti-collision, and deprecation rules.
Agent Tool Pagination Documentation Specification
A documentation specification for paginated tool results consumed by AI agents: cursor schemes, page-size limits, ordering, and termination guarantees.