maintain-integration
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
Chinesemaintain-integration
维护集成包（maintain-integration）
When to use vs create-integration
何时使用 vs create-integration
Use this skill when the package already exists:
- reviewing quality, ECS compliance, or correctness of an existing package
- fixing specific issues: pipeline errors, field mappings, ECS categorization, CEL programs, manifests
- running a full quality improvement pass (review → fix → re-validate loop)
- making minor adjustments to existing data streams
Use instead when:
create-integration- creating a new integration package from scratch
- adding a new data stream to an existing package (→
create-integration)references/add-datastream-workflow.md
当集成包已存在时使用本技能：
- 审核现有集成包的质量、ECS合规性或正确性
- 修复特定问题：管道错误、字段映射、ECS分类、CEL程序、清单
- 执行全面的质量优化流程（审核 → 修复 → 重新验证循环）
- 对现有数据流进行微小调整
在以下场景请改用：
create-integration- 从零开始创建新的集成包
- 向现有集成包添加新数据流（→
create-integration）references/add-datastream-workflow.md
Modes
模式
Review only (read-only, no edits)
仅审核（只读，不编辑）
→ Read fully before starting.
references/review-workflow.mdRun automated validation, delegate inspection to a subagent (see Dispatch convention in ) pointing it at as its operating manual, present findings with no file changes.
references/review-workflow.mdreview-integration/references/reviewer-subagent-guidance.md→ 开始前请完整阅读。
references/review-workflow.md运行自动化验证，将检查任务委托给子代理（请遵循中的调度约定），并指定作为其操作手册，仅呈现检查结果，不修改任何文件。
references/review-workflow.mdreview-integration/references/reviewer-subagent-guidance.mdFull improvement pass (analyze → fix → re-validate)
全面优化流程（分析 → 修复 → 重新验证）
→ Read fully before starting.
references/improve-workflow.mdAnalyze issues (from prior review or fresh reviewer run), prioritize by severity, fix directly or delegate to subagents, re-validate, and report.
→ 开始前请完整阅读。
references/improve-workflow.md分析问题（来自之前的审核或新的审核结果），按严重程度排序，直接修复或委托给子代理，重新验证并提交报告。
Minor direct fix (no subagents needed)
微小直接修复（无需子代理）
For small targeted changes you can handle inline without loading a full workflow:
- manifest field corrections (title, description, format_version, conditions, owner)
- changelog entries and version bumps — see skill
package-spec - documentation placeholder text in
_dev/build/docs/README.md - creation or ECS reference bump
_dev/build/build.yml - simple field file fixes (typos, missing entry, duplicate removal)
- CEL formatting only — run in the stream's
celfmt -s -agent -i cel.yml.hbs -o cel.yml.hbsdirectoryagent/stream/
Run and after any direct edits to confirm no regressions.
elastic-package lintelastic-package check对于无需加载完整流程即可直接处理的小型针对性修改：
- 清单字段修正（标题、描述、format_version、条件、所有者）
- 变更日志条目和版本升级 — 请参考技能
package-spec - 中的文档占位文本
_dev/build/docs/README.md - 创建或升级ECS引用
_dev/build/build.yml - 简单的字段文件修复（拼写错误、缺失条目、重复项移除）
- 仅CEL格式调整 — 在数据流的目录中运行
agent/stream/celfmt -s -agent -i cel.yml.hbs -o cel.yml.hbs
任何直接编辑后，请运行和以确认没有回归问题。
elastic-package lintelastic-package checkSkills to load for direct work
直接工作需加载的技能
- — validation and test commands
elastic-package-cli - — manifest rules, version bumps, and changelog schema
package-spec
Do not load domain-specific skills (pipelines, CEL, ECS, field mappings) into your own context. Delegate to subagents that already have that knowledge.
- — 验证和测试命令
elastic-package-cli - — 清单规则、版本升级和变更日志 schema
package-spec
请勿将领域特定技能（管道、CEL、ECS、字段映射）加载到自身上下文。请委托给已具备相关知识的子代理。
Subagents
子代理
All specialised work is delegated to the platform's generic / general-purpose subagent (Cursor: Task agent; Claude Code: Task agent; or the equivalent on other platforms). Each task prompt must point the subagent at the relevant file by path and instruct it to read that file (plus the skill SKILL.md it lists in "First steps") end-to-end before doing any other work. Do NOT read the guidance file yourself or paste its contents into the task prompt — that doubles its context cost. Pass only the path plus the task-specific context. The subagent will load the manual itself in its own fresh context. Full dispatch rules and per-workflow detail live in and .
generalPurposegeneral-purpose*-subagent-guidance.mdreferences/review-workflow.mdreferences/improve-workflow.md| Subagent guidance file | Use for |
|---|---|
| Thorough read-only quality inspection: classifies files by domain, loads all relevant domain skills and checklists via the |
| Pipeline fixes: JSE00001, error handling, processor tags, ECS categorization, field definitions, test fixtures |
| CEL fixes: program logic, cursor management, error handling, mito validation, mock API, |
When delegating, provide the subagent with: package path, data stream path, specific issues to fix (paste findings), sample data if relevant, and any constraints.
所有专业工作均委托给平台的通用子代理（Cursor：任务代理；Claude Code：任务代理；或其他平台的等效代理）。每个任务提示必须通过路径指向子代理对应的文件，并指示其在开展任何工作前完整阅读该文件（以及文件“第一步”中列出的技能SKILL.md）。请勿自行阅读指导文件或将其内容粘贴到任务提示中——这会加倍上下文成本。仅传递路径和任务特定上下文即可，子代理会在自身全新上下文中自行加载手册。完整的调度规则和各流程细节请参考和。
generalPurposegeneral-purpose*-subagent-guidance.mdreferences/review-workflow.mdreferences/improve-workflow.md| 子代理指导文件 | 适用场景 |
|---|---|
| 全面的只读质量检查：按领域对文件分类，通过 |
| 管道修复：JSE00001错误、错误处理、处理器标签、ECS分类、字段定义、测试用例 |
| CEL修复：程序逻辑、游标管理、错误处理、mito验证、模拟API、 |
委托任务时，请向子代理提供：包路径、数据流路径、需修复的特定问题（粘贴检查结果）、相关样本数据（如有）以及任何约束条件。
Data anonymization
数据匿名化
All data committed must be fully anonymized — no real IPs, hostnames, emails, tokens, or org identifiers in any committed file. When fixing or adding test fixtures, mock responses, sample events, or documentation examples, verify all values are synthetic. Anonymize any real data found as part of the improvement pass.
所有提交的数据必须完全匿名化——提交的任何文件中不得包含真实IP、主机名、邮箱、令牌或组织标识符。修复或添加测试用例、模拟响应、样本事件或文档示例时，请验证所有值均为合成数据。在优化过程中发现的任何真实数据都必须进行匿名化处理。
References
参考资料
- — read-only review workflow (phases 1–4, mandatory checklists, output format)
references/review-workflow.md - — full improvement workflow (analyze → prioritize → fix → re-validate → report)
references/improve-workflow.md
- — 只读审核流程（第1-4阶段、强制检查清单、输出格式）
references/review-workflow.md - — 全面优化流程（分析 → 优先级排序 → 修复 → 重新验证 → 报告）
references/improve-workflow.md