Token导航 LogoToken导航TokenDH.com
开发需要联网github未标认证来源可访问许可证需确认审计通过

define-errors定义错误

Agent Skill

define-errors 用于记录任务执行中的错误、用户纠正、经验和能力缺口,适合在 Codex、Claude、Cursor、Gemini CLI 中希望让 Agent 持续沉淀问题、修正和最佳实践时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

190

周安装

8

GitHub Stars

4,487

下载量

67
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:define-errors(定义错误)
来源仓库:https://github.com/epicenterhq/epicenter
仓库路径:skills/define-errors
安装命令:
npx skills add https://github.com/epicenterhq/epicenter --skill define-errors
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/epicenterhq/epicenter --skill define-errors

简介

用于记录任务执行中的错误、用户纠正和经验缺口,支持持续沉淀问题与最佳实践。

  • 适合在开发流程中修正Agent行为并积累纠错经验时使用。
  • 可与error-handling等技能配合,实现结构化错误处理和类型推断。
  • 安装前应确认权限范围及是否会触发文件读写或命令执行。
  • define-errors 属于开发类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

defineErrors

Related Skills: See error-handling for trySync/tryAsync usage and toast-on-error patterns. See services-layer for service architecture and namespace exports.

When to Apply This Skill

Use this pattern when you need to:

  • Define or refactor domain error variants using defineErrors.
  • Add error variants that include structured fields and cause: unknown.
  • Centralize extractErrorMessage(cause) inside variant factories.
  • Infer union and single-variant types via InferErrors/InferError.
  • Replace old createTaggedError and split Err-pair patterns.

Import

import {
  defineErrors,
  extractErrorMessage,
  type InferErrors,
  type InferError,
} from 'wellcrafted/error';

Core Rules

  1. All variants for a domain live in one defineErrors call — never spread them across multiple calls
  2. The factory function returns {message,...fields} — that is the entire API; no .withMessage(), .withContext(), or .withCause() chains
  3. cause: unknown is just a field like any other — accept it in the input and forward it in the return object
  4. Call extractErrorMessage(cause) inside the factory, never at the call site
  5. Each call like MyError.Variant({...}) returns Err(...) automatically — no separate FooErr pair
  6. Shadow the const with a same-name type using InferErrorsconst FooError / type FooError
  7. Use InferError<typeof FooError.Variant> to extract a single variant's type when needed
  8. Variant names describe the specific failure mode — never use generic names like Service, Error, or Failed
  9. Aim for 2–5 variants per domain, each named by failure mode
  10. Write .message for end-user readabilitytoastOnError shows .message as the muted toast description below the bold title. Write messages that make sense to users, not just developers. Avoid raw paths, status codes, or stack traces as the primary message. Include them after a human-readable prefix:
// ✅ GOOD — human-readable prefix, technical detail after
message: `Could not save recording: ${extractErrorMessage(cause)}`

// ❌ BAD — raw technical output as the entire message
message: `POST /api/recordings 500: ${extractErrorMessage(cause)}`

Patterns

1. Simple variant — no input, static message

export const RecorderError = defineErrors({
  AlreadyRecording: () => ({
    message: 'A recording is already in progress',
  }),
});
export type RecorderError = InferErrors<typeof RecorderError>;

// Call site
return RecorderError.AlreadyRecording();

2. Variant with structured fields — message computed from input

export const DbError = defineErrors({
  NotFound: ({ table, id }: { table: string; id: string }) => ({
    message: `${table} '${id}' not found`,
    table,
    id,
  }),
});
export type DbError = InferErrors<typeof DbError>;

// Call site
return DbError.NotFound({ table: 'users', id: '123' });
// error.message → "users '123' not found"
// error.table   → "users"
// error.id      → "123"

3. Variant with cause — extractErrorMessage inside the factory

import { extractErrorMessage } from 'wellcrafted/error';

export const FfmpegError = defineErrors({
  CompressFailed: ({ cause }: { cause: unknown }) => ({
    message: `Failed to compress audio: ${extractErrorMessage(cause)}`,
    cause,
  }),
  VerifyFailed: ({ cause }: { cause: unknown }) => ({
    message: `Failed to verify temp file: ${extractErrorMessage(cause)}`,
    cause,
  }),
});
export type FfmpegError = InferErrors<typeof FfmpegError>;

// Call site — pass the raw caught error, never call extractErrorMessage here
catch: (error) => FfmpegError.CompressFailed({ cause: error }),

4. Multiple variants in one object — discriminated union built-in

export const DeviceStreamError = defineErrors({
  PermissionDenied: ({ cause }: { cause: unknown }) => ({
    message: `Microphone permission denied. ${extractErrorMessage(cause)}`,
    cause,
  }),
  DeviceConnectionFailed: ({
    deviceId,
    cause,
  }: {
    deviceId: string;
    cause: unknown;
  }) => ({
    message: `Unable to connect to device '${deviceId}'. ${extractErrorMessage(cause)}`,
    deviceId,
    cause,
  }),
  NoDevicesFound: () => ({
    message: "No microphones found. Check your connections and try again.",
  }),
});
export type DeviceStreamError = InferErrors<typeof DeviceStreamError>;
// DeviceStreamError is automatically the union of all three variants

// Extracting a single variant type
type NoDevicesFoundError = InferError<typeof DeviceStreamError.NoDevicesFound>;

5. Domain errors with specific operation failures

export const FsError = defineErrors({
  ReadFailed: ({ path, cause }: { path: string; cause: unknown }) => ({
    message: `Failed to read '${path}': ${extractErrorMessage(cause)}`,
    path,
    cause,
  }),
  WriteFailed: ({ path, cause }: { path: string; cause: unknown }) => ({
    message: `Failed to write '${path}': ${extractErrorMessage(cause)}`,
    path,
    cause,
  }),
  DeleteFailed: ({ path, cause }: { path: string; cause: unknown }) => ({
    message: `Failed to delete '${path}': ${extractErrorMessage(cause)}`,
    path,
    cause,
  }),
});
export type FsError = InferErrors<typeof FsError>;

// Call site
return FsError.ReadFailed({ path: '/tmp/foo.txt', cause: error });

Type Extraction

// Full union type for all variants
type HttpError = InferErrors<typeof HttpError>;

// Single variant type
type ConnectionError = InferError<typeof HttpError.Connection>;

Anti-Patterns

// WRONG — old createTaggedError API
import { createTaggedError } from 'wellcrafted/error';
const { FooError, FooErr } = createTaggedError('FooError')
  .withContext<{ id: string }>()
  .withMessage(({ context }) => `Not found: ${context.id}`);

// WRONG — calling extractErrorMessage at the call site
catch: (error) => MyError.Failed({ message: extractErrorMessage(error) });
// CORRECT — pass raw cause, call extractErrorMessage inside the factory
catch: (error) => MyError.Failed({ cause: error });

// WRONG — one defineErrors per variant (defeats the namespace grouping)
const BusyError = defineErrors({ BusyError: () => ({ message: 'Busy' }) });
const PermError = defineErrors({ PermError: () => ({ message: 'No perm' }) });
// CORRECT — all variants for a domain in one call
const RecorderError = defineErrors({
  Busy: () => ({ message: 'A recording is already in progress' }),
  PermissionDenied: () => ({ message: 'Microphone permission denied' }),
});

// WRONG — using ReturnType instead of InferErrors
type FooError = ReturnType<typeof FooError>;
// CORRECT
type FooError = InferErrors<typeof FooError>;

// WRONG — using separate Err/FooErr pair (old API)
FooErr({ context: { id: '1' } });
// CORRECT — each variant call returns Err(...) automatically
FooError.NotFound({ id: '1' });

// WRONG — generic "Service" variant name (says nothing about the failure mode)
const RecorderError = defineErrors({
  Service: ({ message }: { message: string }) => ({ message }),
});
// RecorderError.Service({ message: '...' }) — "Service" is not a failure mode
// CORRECT — name each variant by what actually went wrong
const RecorderError = defineErrors({
  AlreadyRecording: () => ({ message: 'A recording is already in progress' }),
  PermissionDenied: ({ cause }: { cause: unknown }) => ({
    message: `Microphone permission denied. ${extractErrorMessage(cause)}`,
    cause,
  }),
  DeviceNotFound: ({ deviceId }: { deviceId: string }) => ({
    message: `Device not found: ${deviceId}`,
    deviceId,
  }),
});

// WRONG — generic catch-all with operation string (hides failure modes behind a parameter)
const FfmpegError = defineErrors({
  Service: ({ operation, cause }: { operation: string; cause: unknown }) => ({
    message: `Failed to ${operation}: ${extractErrorMessage(cause)}`,
    operation,
    cause,
  }),
});
// FfmpegError.Service({ operation: 'compress audio', cause }) — variant name is meaningless
// CORRECT — each operation is its own variant
const FfmpegError = defineErrors({
  CompressFailed: ({ cause }: { cause: unknown }) => ({
    message: `Failed to compress audio: ${extractErrorMessage(cause)}`,
    cause,
  }),
  VerifyFailed: ({ cause }: { cause: unknown }) => ({
    message: `Failed to verify temp file: ${extractErrorMessage(cause)}`,
    cause,
  }),
});

// WRONG — monolithic single-variant error for a domain with many failure modes
const RecorderError = defineErrors({
  Error: ({ message }: { message: string }) => ({ message }), // Too vague
});
// CORRECT — split by failure mode
const RecorderError = defineErrors({
  AlreadyRecording: () => ({ message: 'A recording is already in progress' }),
  InitFailed: ({ cause }: { cause: unknown }) => ({
    message: `Failed to initialize recorder: ${extractErrorMessage(cause)}`,
    cause,
  }),
  StreamAcquisition: ({ cause }: { cause: unknown }) => ({
    message: `Failed to acquire recording stream: ${extractErrorMessage(cause)}`,
    cause,
  }),
});

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.6%
按下载量换算23

Claude

30.72%
按下载量换算21

Cursor

20.22%
按下载量换算14

Gemini CLI

8.71%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills