review

Prompt file imported from ZVanoZ/ts-react-petrichenko (.cursor/commands/review.md). Copyright stays with the author.

REVIEW Command - Course Material & Solution Analysis

This command focuses on the @TODO: "Cursor AI" block from the root readme.md. It checks lesson relevance using up‑to‑date documentation (via MCP user-context7 / @context7) and attaches short results to the course table of contents, while writing detailed reports next to each lesson.

The command is independent from vanzan01 "cursor-memory-bank": it does not require any Memory Bank files to exist and does not create or update them. It works only with lesson folders and the project documentation files described below.

Responsibilities

The /review command implements the following high‑level tasks:

1. Relevance check for a single lesson

   • Use up‑to‑date documentation and examples (via MCP user-context7 / @context7) to verify that the lesson content is still accurate and complete.
   • Write a detailed report to ai-review.md in the lesson folder.
   • Add or update a collapsible details block in the root readme.md under the corresponding lesson entry in the ## Содержание section.

2. Bulk relevance check for all lessons

   • Iterate over all lessons listed in the ## Содержание section of the root readme.md.
   • For each lesson, perform the same workflow as for a single lesson, updating both the per‑lesson ai-review.md and the relevant details block in readme.md.

3. Solution quality analysis

   • For a given lesson, analyze the user's solution:
     • highlight knowledge gaps;
     • identify weak or non‑idiomatic implementations;
     • recommend concrete improvements and next study steps.
   • Record this analysis in ai-review.md (in a dedicated section) or in a separate lesson-review section, depending on the lesson structure.

4. Usage guide for Cursor AI

   • Generate and maintain a concise guide that explains how to use Cursor AI effectively in this repository (including /van, /plan, /build, /review and other commands).
   • Store this guide in cursor-ai-usage.md at the repository root and reference it from the root readme.md.

Subcommands

/review lesson <lesson-path>

Purpose: Run a relevance check for a single lesson and update its summary in the root readme.md.

• Input
  • lesson-path: relative path to the lesson directory (for example, 20-ts/40-classes/110-this-and-typed-context).
• Expected lesson content (typical)
  • readme.md with theory, tasks or notes and/or
  • src directory with the lesson source code.
• Workflow
  1. Determine lesson context
     • Read readme.md and/or relevant files under lesson-path (especially src).
     • Extract key topics and concepts to verify (for example, Type aliases, Generics functions, this typing, etc.).
  2. Check relevance via MCP user-context7 / @context7
     • For each key topic, query the external documentation and examples:
       • identify breaking changes, deprecated patterns, or new best practices;
       • detect missing but important subtopics.
  3. Write or update lesson-path/ai-review.md
     • Create the file if it does not exist.
     • At minimum, include:
       • date of the check;
       • overall status;
       • detailed per‑topic notes;
       • recommended extra topics to learn if content is incomplete or outdated.
  4. Update the corresponding entry in the root readme.md
     • Locate the lesson item in the ## Содержание section:
       • find the numbered list entry describing the lesson;
       • use the immediately following bullet with a link (for example, * [src](20-ts/40-classes/110-this-and-typed-context)) to match lesson-path.
     • Under this entry, ensure there is exactly one collapsible block:

       <details>
           <summary>ai-relevance</summary>
       
           [ai-review.md](<lesson-path>/ai-review.md)
           ```text
           Дата проверки: <YYYY-MM-DD>
           Статус: <Актуально | Частично устарело | Устарело>
           ```
         </details>

     • If the block does not exist, create it in this exact format, preserving the surrounding indentation style.
     • If the block already exists, only update:
       • the link to ai-review.md (if the path has changed);
       • the Дата проверки line;
       • the Статус line.

/review all

Purpose: Perform relevance checks for all lessons listed in the ## Содержание section.

• Workflow
  1. Read the root readme.md.
  2. Parse the ## Содержание section and collect all lesson entries that:
     • have a numbered item with a lesson title;
     • are immediately followed by a bullet item containing a link either to src or to readme.md in a lesson directory (for example, * [src](20-ts/40-classes/110-this-and-typed-context)).
  3. For each collected lesson:
     • determine lesson-path from the link target;
     • run the same workflow as /review lesson <lesson-path>:
       • compute relevance using up‑to‑date docs;
       • write/update ai-relevance.md;
       • add/update the corresponding <details> block under that lesson entry in readme.md.
  4. Optionally, implementations may support skipping recently checked lessons (for example, based on Дата проверки within a certain threshold), but this is an optimization, not a requirement of the command contract.

/review analyze <lesson-path>

Purpose: Analyze the quality of the user's solution for a specific lesson and point out knowledge gaps and weak parts of the implementation.

• Input
  • lesson-path: relative path to the lesson directory where the user's work resides.
• Workflow
  1. Read the solution code and any solution‑specific readme.md under lesson-path.
  2. Using MCP user-context7 / @context7 and, where available, the official or reference solutions:
     • compare interfaces, types, patterns and architecture;
     • identify:
       • missing required features for this lesson;
       • non‑idiomatic or unsafe TypeScript or React usage;
       • unnecessary complexity or code smells.
  3. Write analysis into lesson-path/ai-review.md:
     • Ensure there is a single top‑level section ## Анализ вашего решения that acts as a chronological log of all analysis runs for this lesson.
     • For each invocation of /review analyze <lesson-path>:
       • insert a new sub‑section at the top of ## Анализ вашего решения with the heading:
         • ### Дата: <YYYY-MM-DD HH:mm> (24‑hour format, where HH = hours and mm = minutes, using the actual local analysis time);
       • inside this sub‑section, structure the content using the following blocks and numbered lists:

         ## Анализ вашего решения
         
         ### Дата: <YYYY-MM-DD HH:mm>
         
         #### Выполнено:
         
         1. "<OLD-DATE> п.N"
         
         #### "@TODO:
         
         1. "<OLD-DATE> п.M"
         
         #### Пробелы в знаниях
         
         1. ...
         
         #### Замечания к реализации
         
         1. ...
         
         #### Рекомендованные шаги
         
         1. ...

         where:
         • "<OLD-DATE> п.N" is a reference to item N from an earlier ### Дата: ... section in the same ai-review.md file.
     • When constructing Выполнено and "@TODO:
       • parse all earlier ### Дата: ... sections within ## Анализ вашего решения;
       • for each previously suggested item, determine whether the current solution code now satisfies it:
         • if yes, add a reference under Выполнено and treat this recommendation as closed;
         • if no, add or keep a reference under "@TODO until it is eventually satisfied in a future run.
     • Preserve all previous ### Дата: ... sections verbatim to maintain a full history; never overwrite or delete historical analysis entries.
  4. Keep the short <details> block in readme.md focused on the overall status and date; deep analysis remains only in ai-review.md.

AI markers in code (@ai-review)

To help /review analyze associate concrete pieces of code with specific recommendations in ai-review.md, you can place markers in comments inside lesson code and related scripts.

• Generic marker syntax (human‑readable path)

  @ai-review|Анализ вашего решения/Дата: "<YYYY-MM-DD HH:mm>"/<block-title>/<item-number>|<free-text>

  Where:

  • The part between @ai-review| and the final | is a path that mirrors the physical structure of ai-review.md:
    • Анализ вашего решения – the ## Анализ вашего решения section.
    • Дата: "<YYYY-MM-DD HH:mm>" – the exact ### Дата: ... heading for this analysis run.
    • <block-title> – one of the subsection headings under that date, for example:
      • Пробелы в знаниях
      • Замечания к реализации
      • Рекомендованные шаги
    • <item-number> – 1‑based index of the bullet within that subsection’s numbered list.
  • <free-text> – optional human‑readable explanation for maintainers; it does not participate in linking logic.

  When scanning code, /review analyze should:

  • Look for lines containing @ai-review|.
  • Split the marker on the first | into:
    • the path (Анализ вашего решения/.../<item-number>);
    • optional <free-text>.
  • Use the path to locate the referenced recommendation inside ai-review.md:
    • find ## Анализ вашего решения;
    • within it, the matching ### Дата: "<YYYY-MM-DD HH:mm>";
    • within that date block, the subsection with title <block-title>;
    • and then the <item-number>‑th item in that subsection’s list.

  This linkage can then be used when deciding whether a specific recommendation should be moved from "@TODO to Выполнено in a new analysis entry.

/review cource-plan-for-ai

Purpose: Refresh the AI-facing course plan used by /review based on the current source URL.

• Workflow
  1. Read cource-plan-for-ai.md at the repository root.
  2. Read the ## "ai-options" section and parse the ini code block.
  3. Extract plan-url from the ini block (for example: plan-url="https://...").
  4. Perform an HTTP request to plan-url and, исходя исключительно из структуры страницы курса (а не локального readme.md), построить иерархический набросок плана:
     • разделы и подмодули курса;
     • отдельные уроки с их номерами и ссылками.
  5. Validation (mandatory): verify that the fetched page actually contains the lesson list:
     • if the page is paywalled / requires authentication / does not contain a parseable plan, the command must:
       • stop;
       • write a short note in cource-plan-for-ai.md (above ## "cource-plan") that the plan cannot be regenerated automatically from plan-url and why;
       • NOT regenerate ## "cource-plan" from any local sources by default.
     • The command must never silently fall back to using the local readme.md as the source of truth.
  6. Optional fallback (explicit only):
     • implementations may support a flag like --from-readme to regenerate the plan from the local readme.md, but only when the user explicitly requests it.
  7. Regenerate the ## "cource-plan" section in cource-plan-for-ai.md:
     • replace everything under ## "cource-plan" with a hierarchical markdown bullet list that mirrors the course page structure (sections, subsections, lessons with their links and, при наличии, путями к локальным директориям);
     • keep the introductory part of the file (description, rules) intact.
  8. The regenerated plan must preserve lesson order so that /review lesson and /review analyze can:
     • know which lessons come before and after the current one;
     • avoid marking topics as "not covered" in ai-relevance.md if those topics belong to lessons that come after the current lesson in the plan.

/review guide

Purpose: Create or update a short guide on how to use Cursor AI and the project‑specific commands effectively.

• Workflow
  1. Create or update cursor-ai-usage.md at the repository root.
  2. Include at least:
     • how to call /van, /plan, /build, /review;
     • how the Memory Bank system is used conceptually (without making /review depend on it);
     • how relevance checks and solution analysis work at a high level;
     • tips on requesting concise answers and avoiding unnecessary context to reduce token usage.
  3. Ensure that the root readme.md references cursor-ai-usage.md from the ### Помощник AI section or another clearly visible place.

Status Values

The Статус line in the ai-relevance block and in ai-review.md should use one of the following values:

• Актуально
• Частично устарело
• Устарело

Implementations may also provide additional explanation or sub‑status inside ai-relevance.md, but these three values form the stable external contract exposed in the root readme.md.

Usage

• /review lesson 20-ts/40-classes/110-this-and-typed-context
  Check the relevance of the this и типизация контекста lesson, write/update ai-relevance.md in its folder and attach/refresh the ai-relevance details block under this lesson in the ## Содержание section.

• /review all
  Iterate through all lessons listed in the ## Содержание section of readme.md, perform relevance checks, and update their associated ai-relevance blocks.

• /review analyze 20-ts/30-generics-manipulation/160-practice
  Analyze the user's solution for the generics practice lesson and record gaps and issues in ai-relevance.md.

• /review cource-plan-for-ai
  Read cource-plan-for-ai.md, take plan-url from the ## "ai-options" ini block, fetch the actual course plan, and fully regenerate the ## "cource-plan" section so that subsequent /review calls can correctly distinguish between current and future topics.

• /review guide
  Generate or refresh the cursor-ai-usage.md guide and ensure it is referenced from the root readme.md.