القائمة
System design, task decomposition, and technical planning. Use when designing new features, planning refactors, breaking down large tasks, or creating technical specifications.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | Architecture & Planning |
| description | System design, task decomposition, and technical planning. Use when designing new features, planning refactors, breaking down large tasks, or creating technical specifications. |
You are creating a technical plan. Be concrete and actionable — avoid vague hand-waving.
This skill should always externalize the work as a todo plan before deeper exploration. Do not keep the plan only in prose.
read — read existing code, configs, schemas to understand current architecturesearch — find patterns, dependencies, callers, implementationsexecute — check project structure, dependency tree, git historyweb — research libraries, patterns, comparable implementationstodo — track planning phases and deliverablesagent — delegate research sub-tasks to helper agentstodo({ todoList: [
{ id: 1, title: "Map existing architecture", status: "in-progress" },
{ id: 2, title: "Explore design options", status: "not-started" },
{ id: 3, title: "Select approach and justify", status: "not-started" },
{ id: 4, title: "Decompose into tasks", status: "not-started" },
{ id: 5, title: "Identify risks and open questions", status: "not-started" }
]})
Use read and search to understand the current codebase:
search to map how components connectexecute to check project structure and dependenciesIdentify 2-3 viable approaches. For each, note:
web to research relevant libraries or prior artChoose the approach with the best trade-off profile. Justify why.
Break into ordered, concrete implementation tasks:
todo as you complete each planning phasetodo as mandatory for this skillProve bug fixes with fail-then-pass validation. Use when debugging regressions, user-reported bugs, flaky behavior, race conditions, UI stream issues, or any fix where the user asks to reproduce the issue first, validate the fix, avoid masking symptoms, or show that the old behavior fails and the new behavior passes.
Systematic code review covering correctness, security, performance, and maintainability. Use when reviewing pull requests, auditing code changes, or assessing code quality.
Structured data exploration, transformation, and analysis. Use when working with datasets, CSV/JSON processing, database queries, data visualization, or ETL pipelines.
Systematic root cause analysis and bug fixing. Use when diagnosing errors, crashes, unexpected behavior, test failures, or any "it's not working" scenario.
Systematic web research with source triangulation, structured synthesis, and citations. Use when the task requires investigating topics, comparing options, finding documentation, or answering questions that need external knowledge.
Performance analysis, profiling, and optimization. Use when diagnosing slow code, optimizing database queries, reducing memory usage, or improving application responsiveness.