AI 智能体
agent-guardrails-error-handling
更新于 2026-08-06
Agent Guardrails · 错误处理规则
Error-handling rules for Agent Guardrails: classification, typed codes and single-point mapping to user outcomes.
- 技术栈
- Agent Guardrails
- 主题
- 错误处理
- 条目
- 9 条
- 版本
- v0.2.0
✓ 应当这么做
- ✓ Classify failures once: expected (validation, not-found) vs unexpected (bugs, dependencies down)
- ✓ Return typed errors carrying a stable code plus human-readable context
- ✓ Map errors to user-visible outcomes at the outermost layer only
✕ 常见反模式
- ✕ String matching on error messages to decide control flow
- ✕ Swallowing an error to keep the happy path green
- ✕ Leaking stack traces or internal identifiers to end users
◎ 交付前验证
- ◎ Injection cases added to the adversarial test corpus
- ◎ Tool allow-lists reviewed; new tools default to deny
- ◎ Secrets scanned in output before responding
同分类条目
42 篇AI 智能体
Vector Search 工程规则
Vector search rules: chunking strategy, recall evaluation and index rebuild path
AI 智能体
Feature Store 工程规则
Feature store rules: point-in-time correctness and train/serve parity
AI 智能体
Agent Skill Design 工程规则
Skill design rules: narrow triggers, explicit boundaries and verifiable outputs
AI 智能体
Agent Tool Calls 工程规则
Tool-call rules: server-side authorisation and deny-by-default allow-lists
AI 智能体
Agent Memory 工程规则
Memory rules: scoped retrieval, expiry and user-visible deletion
AI 智能体
RAG Retrieval 工程规则
RAG rules: citation-grade retrieval and evaluated recall before shipping