RCP-CWK-019 Cowork Cross-Cookbook Scanner

A complete picture of every recipe across all your cookbooks

Scans all CRAFT cookbook files in your workspace, consolidates recipe inventories across cookbooks, generates unified cheatsheets, and performs cross-cookbook analysis including Core-to-Cowork extension mappings.


Cowork Cross-Cookbook Consolidated Scanner

Tags: Introduced in Beta, Cookbook, Scanner, Cheatsheet, Inventory, Cross-Cookbook, Discovery, Bridge, Cowork

TL;DR

What It Does
Scans all CRAFT cookbook files in your workspace, consolidates recipe inventories across cookbooks, generates unified cheatsheets, and performs cross-cookbook analysis including Core-to-Cowork extension mappings.
How It Works
The recipe uses Glob to discover all CFT-FWK-COOKBK-*.txt files in the project folder, then reads each cookbook and detects recipe markers to extract metadata: ID, title, difficulty, category, version, description, parameters, and extension relationships. It consolidates everything into a unified inventory, analyzes cross-cookbook relationships (extension chains, dependency mappings, duplicates), and generates output in your chosen format — summary table, detailed breakdown, Core-vs-Cowork comparison, or exported file.
What To Expect
A complete picture of every recipe across all your cookbooks. In summary format you get a compact table. In detailed format you get full metadata per recipe. In comparison format you see how each Cowork recipe extends its Core ancestor. In export format you get a saved cheatsheet file ready for git checkpoint.
Best Results When You
Run with scope “all” and format “summary” first to get the big picture, then drill into specific cookbooks or categories. Use “comparison” format when you want to understand how Cowork recipes build on Core. Use “export” to create a persistent cheatsheet you can reference across sessions.
Typical Time
Under 1 minute for summary format. 1–2 minutes for detailed or export formats that process all recipe metadata.
Difficulty
Easy — the scanner does all the work automatically. Just choose a scope and output format. No parameters are required; defaults scan all cookbooks in summary format.

How To Start

The scanner discovers cookbooks automatically. Choose your scope and output format, or accept the defaults.

STEP 1Configure the Scan

Tell the AI what to scan and how to present the results. All parameters are optional — defaults scan all cookbooks in summary format.
Available parameters
  • scope · string · optional · default: all
    Scan scope. Options: “all” (scan all discovered cookbooks), “single” (scan one specific cookbook file), “category” (scan by cookbook category), “search” (keyword search across all cookbooks).
  • cookbook_file · string · optional
    Specific cookbook filename to scan. Used with scope “single.”
  • search_term · string · optional
    Keyword to search across all cookbooks. Used with scope “search.”
  • output_format · string · optional · default: summary
    How to present results. Options: “summary” (compact table), “detailed” (full metadata per recipe), “comparison” (Core-to-Cowork side-by-side), “export” (write cheatsheet to file).
Invocations
Scan All (Default)
#H->AI::Directive: (Scan cookbooks and generate consolidated cheatsheet)
Scan Single Cookbook
#H->AI::Directive: (Scan cookbooks and generate consolidated cheatsheet with scope=single, cookbook_file=“CFT-FWK-COOKBK-COWORK.txt”)
Keyword Search
#H->AI::Directive: (Scan cookbooks and generate consolidated cheatsheet with scope=search, search_term=“handoff”)
Core vs Cowork Comparison
#H->AI::Directive: (Scan cookbooks and generate consolidated cheatsheet with output_format=comparison)
Export Cheatsheet
#H->AI::Directive: (Scan cookbooks and generate consolidated cheatsheet with output_format=export)

Usage Examples

Three scenarios — getting a quick inventory, comparing Core and Cowork recipes, and searching for a specific capability.
Example 1 · Quick inventory of all recipes
User: How many recipes do I have across all cookbooks?
AI: Scanning workspace. Found 3 cookbook files. Scanning recipe markers. Scan complete. Cookbooks scanned: 3. Total recipes found: 128. By cookbook: CORE 22, COWORK 48, COWORK-ADMIN 58. Orphaned markers: 0. Cross-cookbook extensions: 21. Here is the summary table.
Example 2 · Core-to-Cowork comparison
User: Show me how Cowork recipes extend Core.
AI: Generating comparison. Found 21 Cowork recipes that extend Core ancestors. CORE: Token Monitor → COWORK: Token Monitor with File-Based Logging. Additions: persistent logging, Cowork-specific capacity estimates, LATE_SESSION_MODE flag. [continues for all 21 mappings]
Example 3 · Search for handoff-related recipes
User: Which recipes deal with session handoffs?
AI: Searching all cookbooks for “handoff.” Found 4 matches: RCP-001-001-002 Session Handoff (Core), RCP-CWK-002 Cowork Session Handoff (Cowork), RCP-CWK-ADM-070 Global File Manager (Admin — references handoff), RCP-CWK-010 Token Monitor (Cowork — auto-handoff trigger). Would you like detailed info on any of these?

How AI Reads This Recipe

When this recipe is triggered, the AI scans cookbooks and generates consolidated recipe inventories through a 6-step pipeline. The AI should:
  1. DISCOVER COOKBOOKS by scanning the project folder with Glob for CFT-FWK-COOKBK-*.txt files. Validate the discovered files against the expected cookbook list from CLAUDE.md. Report any discrepancies.
  2. FILTER BY SCOPE: “all” scans everything, “single” targets one cookbook file, “category” targets a cookbook category, “search” uses Grep to find keyword matches across all cookbooks.
  3. DETECT RECIPE MARKERS in each cookbook using two patterns: “# START RECIPE-ID: RCP-” and “# RECIPE: RCP-”. Extract 9 metadata fields from each recipe: recipe_id, title, difficulty, category, version, description (summarized to 2 sentences), parameters, extends, and dependencies. Track orphaned markers (start without matching end).
  4. ANALYZE CROSS-COOKBOOK relationships: total recipe counts, distribution by category, Core-to-Cowork extension mappings (via extends field), cross-cookbook dependency chains, duplicate or overlapping IDs, and orphaned markers.
  5. GENERATE OUTPUT in the requested format: summary (compact table), detailed (full metadata per recipe per cookbook), comparison (Core recipe → Cowork adaptation → delta summary), or export (written to project folder as a dated cheatsheet file with git checkpoint recommendation).
  6. DISPLAY SUMMARY with cookbooks scanned, total recipes, breakdown by cookbook, orphan count, and cross-cookbook extension count. Offer follow-up options: search, detail, export, or re-scan.
The recipe extends Core Recipe Cheatsheet with filesystem-native multi-cookbook scanning and cross-cookbook analysis. Never fabricate recipe metadata — always read from the actual cookbook file. If a cookbook file is unreadable, skip it and report the error. The comparison format is particularly useful for understanding how the Cowork cookbook builds on Core foundations.

When to Use This Recipe

Use this recipe when you:
  • Want a complete inventory of all recipes across all cookbooks in a single consolidated view.
  • Need to understand how Cowork recipes extend their Core ancestors via comparison format.
  • Are searching for a recipe by keyword and do not know which cookbook it is in.
  • Want to export a persistent cheatsheet file for cross-session reference.
Do not use this recipe when:
You already know the exact recipe ID and just want to run it. Use CWK-018 (Recipe Loader) instead for direct load and execute. The scanner is for discovery and inventory — not execution.

Recipe FAQ

Q.What cookbook files does the scanner look for?

Any file matching the pattern CFT-FWK-COOKBK-*.txt in the project folder. The active cookbooks are CORE (22 recipes), COWORK (48 recipes), and COWORK-ADMIN (58 recipes, across 8 volumes). The scanner validates discovered files against the expected list in CLAUDE.md and reports discrepancies.

Q.How are recipe markers detected?

Two patterns: “# START RECIPE-ID: RCP-” and “# RECIPE: RCP-”. The scanner looks for both because cookbook formatting varies. Each start marker should have a matching end marker. Start markers without ends are flagged as orphaned.

Q.What metadata is extracted from each recipe?

Nine fields: recipe_id, title, difficulty, category, version, description (AI-summarized to 2 sentences), parameters (with required/optional flags), extends (Core ancestor if applicable), and dependencies from the dependency map.

Q.What does the comparison format show?

A side-by-side view of each Core recipe and its Cowork adaptation. For each pair it shows: Core recipe ID → Cowork recipe ID → what the Cowork version adds (persistent logging, file-based state, new parameters, Cowork-specific integrations, etc.).

Q.Can I scan archived cookbooks?

The scanner discovers cookbooks by file pattern in the project folder. Archived cookbooks (like the former STUDIO and ADMIN catalogs) are stored in framework/archive/ and will not appear in the scan unless you specify their file path with scope “single.”

Q.How does export format work?

It writes a complete cheatsheet to the project folder as CRAFT-RECIPE-CHEATSHEET-{date}.txt with full metadata and execution codes for every recipe. The recipe recommends a git checkpoint after export for version persistence.

Version History

Changes to this recipe over time. Most recent first.
v1.00a 2026-02-28
Initial release. Extends Core Recipe Cheatsheet with Cowork filesystem-native scanning: 4 scope options (all, single, category, search), 4 output formats (summary, detailed, comparison, export), multi-cookbook discovery via Glob, dual recipe marker detection, 9-field metadata extraction, cross-cookbook analysis (extension mappings, dependency chains, duplicate detection, orphan tracking), file export with git checkpoint recommendation.

Get this recipe with CRAFT for Claude Cowork

Cowork recipes ship bundled with CRAFT for Claude Cowork — there’s no separate download. Clone the framework once, and your AI runs every recipe automatically when invoked.

Pull anytime to stay on the latest version — free to clone, no login or email required.

Then start your session

Once CRAFT is in your project folder, open a new Cowork session and ask Claude to initialize. For example:

You

Please initialize my CRAFT session.

Claude

CRAFT session ready. Your project is loaded, your persona is active, and your recipes are available. What would you like to work on?

What is CRAFT for Claude Cowork?

Not familiar with Git? Download as a ZIP

No command line needed. Just download, move, and unzip:

  1. Open the CRAFT framework repo on GitHub.
  2. Click the green Code button, then choose Download ZIP.
  3. Move the downloaded ZIP into your Claude Cowork project folder.
  4. Unzip it: double-click on Mac, or right-click → Extract All on Windows.

Similar Posts