wb-flow 1.0.0-r01

package/templates/shortcuts/shortcuts.md

@@ -0,0 +1,413 @@
+# My AI Shortcuts v4.3 - Universal Control Layer
+Last updated: April 2026
+Works with: Grok, Claude, ChatGPT, Gemini, Cursor, Antigravity, any LLM
+
+---
+
+## UNIVERSAL PREFIX (paste this at the start of every new chat)
+
+You are now using my personal AI shortcut system v4.3.
+From now on and until I say "/ignore_previous", interpret every shortcut I use exactly as defined in this file. Never explain the shortcuts unless I ask "/help". Never ignore them.
+
+DYNAMIC ARGUMENT RULES:
+- Basic: /shortcut
+- With args: /shortcut,arg1,arg2 (no spaces)
+- Aliases: /ctx=/folder, /f=/file, /p=/project, /ep=/prompt
+- Chain freely: /d /ctx,foldA,refactor /ep,optimize /final_only
+
+PREFIX ALIAS RULE:
+Every shortcut may be invoked with either `/` or `!` as the prefix — they are fully interchangeable (`!quote` == `/quote`, `!c` == `/c`). Use `!` when `/` would collide with a CLI/host command (e.g., `!quote` to avoid the host's `/quote` parser). All rules, args, and chaining behave identically regardless of prefix.
+
+PRIORITY RULES:
+1. Natural language overrides shortcuts
+2. Later overrides earlier
+3. Specific overrides general
+4. /final_only and /answer_only override all formatting
+
+CONTEXT PROTOCOL:
+When I paste code or text after a /ctx, /f, or /p command, treat it as the content for that named slot.
+Confirm with: "✓ Context [name] loaded — [one-line summary]"
+
+Default style: friendly, direct, zero fluff.
+
+---
+
+## CORE MODE (use for 90% of tasks)
+/c /d /vd
+/b /table /step
+/final_only /no_explanation
+/first_principles /explain_reasoning /cot
+/ask /checklist
+/ctx,foldA,load → then paste content
+/ctx,foldA,analyze /f,file.py,debug
+/ep,optimize
+/execute /goal,build_api
+
+---
+
+## 1. STYLE & LENGTH
+
+/c /concise          → Short, essential only
+/d /detailed         → Full coverage with context
+/vd /verydetailed    → Exhaustive, include edge cases
+/length,short        → ~50 words
+/length,N            → ~N words (e.g. /length,200)
+/summary             → Condensed version of the topic or output
+/deep_dive           → Go as deep as possible
+/eli5                → Explain like I'm 5
+/technical           → Use precise technical terminology
+/non_technical       → Plain language, no jargon
+/practical           → Focus on real-world application
+/theoretical         → Focus on concepts and principles
+
+---
+
+## 2. TONE
+
+/wissem      → Direct, technical, no filler. Aware of Tunisian/startup context.
+/casual      → Relaxed, conversational
+/pro         → Formal and professional
+/direct      → No hedging, no fluff, get to the point
+/empathetic  → Warm, understanding tone
+/sarcastic   → Dry wit (use carefully)
+/optimist    → Emphasize possibilities and upsides
+/pessimist   → Emphasize risks and downsides
+/motivational → Encouraging and energizing
+/neutral     → Balanced, no opinion
+/persuasive  → Make the strongest case
+
+---
+
+## 3. FORMAT
+
+/b /bullet       → Bullet list
+/nb /numbered    → Numbered list
+/step            → Step-by-step instructions
+/table           → Structured table
+/json            → JSON format
+/yaml            → YAML format
+/csv             → CSV format
+/mindmap         → Text-based mind map
+/faq             → Question & answer format
+/proscons        → Pros and cons layout
+/timeline        → Chronological structure
+/compare,a,b     → Side-by-side comparison of a vs b
+/code            → Code block output
+/md              → Raw markdown output
+/blueprint       → Actionable, step-by-step technical execution guide/specification
+
+---
+
+## 4. OUTPUT CONTROL
+
+/final_only      → Only the final output, no process or commentary
+/answer_only     → Answer only, no intro or context
+/no_explanation  → Skip all explanations
+
+---
+
+## 5. REASONING
+
+/explain_reasoning  → Show the reasoning behind the answer
+/cot                → Chain of Thought: think step by step before answering
+/tot                → Tree of Thought: explore multiple reasoning paths, then conclude
+/selfcritic         → Generate output, then critique it, then improve it
+/first_principles   → Break down to fundamentals before reasoning up
+/rootcause          → Identify the root cause, not just symptoms
+/swot               → Structure as Strengths, Weaknesses, Opportunities, Threats
+/analogy            → Explain using a clear real-world analogy
+
+---
+
+## 6. REFINEMENT
+
+/improve    → Make it better overall
+/simplify   → Reduce complexity
+/expand     → Add more depth and detail
+/shorten    → Make it more concise
+/rewrite    → Fully rewrite from scratch
+
+---
+
+## 7. UNCERTAINTY
+
+/assumptions  → List all assumptions being made
+/unknowns     → Flag what is unknown or unclear
+/confidence   → State confidence level for each claim
+
+---
+
+## 8. EXAMPLES
+
+/examples,N       → Generate N examples (e.g. /examples,3)
+/examples,real    → Use real-world examples only
+/examples,simple  → Use simple, beginner-friendly examples
+/examples,code    → Use code examples
+
+---
+
+## 9. INTERACTION
+
+/ask          → Ask me clarifying questions before proceeding
+/no_questions → Do not ask questions, make best assumptions
+/interactive  → Turn-by-turn mode, wait for my input at each step
+/steps,N      → Break output into N steps
+/phases       → Organize output into phases
+/checklist    → Output as an actionable checklist
+/confirm      → Ask me to confirm before taking any action
+/iterate,N    → Refine the output N times before presenting final version
+
+---
+
+## 10. DEBUG
+
+/debug          → Identify what is wrong and why
+/fix            → Fix the issue and explain the change
+/validate       → Check correctness and flag issues
+/explain_error  → Explain the error in plain language
+/trace          → Trace execution flow step by step
+/reproduce      → Show minimal code to reproduce the issue
+
+---
+
+## 11. ROLES
+
+/expert      → World-class expert in the relevant domain
+/teacher     → Explain clearly for learning
+/professor   → Academic depth and rigor
+/coach       → Practical guidance and motivation
+/critic      → Honest critical evaluation
+/ceo         → Strategic, high-level business thinking
+/scientist   → Data-driven, hypothesis-based reasoning
+/engineer    → Implementation-focused, pragmatic
+/detective   → Investigate, find inconsistencies, draw conclusions
+/storyteller → Narrative and engaging communication style
+/lawyer      → Precise, risk-aware, cover all angles
+/designer    → UX/UI and visual thinking perspective
+/pm          → Product manager: prioritize, scope, ship
+
+---
+
+## 12. LANGUAGE
+
+/en               → Respond in English
+/fr               → Respond in French
+/ar               → Respond in Arabic
+/tn               → Respond in Tunisian dialect (Darija)
+/translate,en,fr  → Translate input from English to French
+/bilingual,en,fr  → Output in both English and French
+
+---
+
+## 13. CREATIVE
+
+/creative       → Creative approach, think outside the box
+/very_creative  → Maximum creative freedom
+/brainstorm,N   → Generate N distinct ideas
+/innovate       → Push for novel, non-obvious solutions
+/fiction        → Creative fiction mode
+/analogy        → (see Section 5 — also usable standalone)
+
+---
+
+## 14. PROMPT OPERATIONS
+
+/ep              → Analyze and optimize the current prompt
+/ep,analyze      → Diagnose weaknesses in the prompt
+/ep,optimize     → Rewrite for better AI performance
+/ep,expand       → Add more detail and context to the prompt
+/ep,compress     → Make the prompt shorter while keeping intent
+/ep,structure    → Organize the prompt for clarity
+/ep,precise      → Make the prompt more specific and unambiguous
+/ep,robust       → Make the prompt handle edge cases
+/ep,chain        → Turn prompt into a multi-step chain
+/ep,role,expert  → Add a role definition to the prompt
+/ep,format,json  → Specify JSON output format in the prompt
+/ep,system       → Rewrite as a system prompt
+/ep,agent        → Rewrite as an agentic prompt
+/ep,multi_agent  → Decompose into a multi-agent workflow
+/ep,tooling      → Add tool-use or function-calling structure
+/ep,interactive  → Make prompt interactive / turn-based
+/ep,evaluate     → Add an evaluation criterion to the prompt
+/ep,test         → Generate test cases for this prompt
+
+---
+
+## 15. CONTEXT & FILE MANAGEMENT
+
+Aliases: /ctx=/folder, /f=/file, /p=/project
+Hierarchy: project > folder > file
+
+### Load / Set Context
+/ctx,foldA,load              → Paste content after this; AI confirms load with one-line summary
+/ctx,foldA,set,<description> → Define what foldA contains without pasting full content
+/ctx,foldA,scope,<area>      → Narrow focus within the folder (e.g. scope,auth_module)
+/ctx,foldA,reset             → Clear this context slot
+/ctx,list                    → List all currently active context slots
+
+/f,file.py,load              → Paste file content; AI confirms and summarizes
+/f,file.py,set,<description> → Define what file.py is without pasting it
+
+### Folder Actions
+/ctx,foldA,overview      → High-level one-paragraph summary
+/ctx,foldA,analyze       → Deep structural and logic analysis
+/ctx,foldA,document      → Generate documentation
+/ctx,foldA,improve       → Suggest improvements
+/ctx,foldA,refactor      → Refactor for clarity and maintainability
+/ctx,foldA,debug         → Find bugs and issues
+/ctx,foldA,secure        → Identify security vulnerabilities
+/ctx,foldA,optimize      → Optimize for performance
+/ctx,foldA,test          → Generate test cases
+/ctx,foldA,map           → Visual map of structure and relationships
+/ctx,foldA,explain       → Plain-language explanation of what it does
+/ctx,foldA,dependencies  → List internal and external dependencies
+/ctx,foldA,entrypoints   → Identify all entry points
+/ctx,foldA,architecture  → Describe architectural pattern
+/ctx,foldA,dataflow      → Trace data flow through the folder
+/ctx,foldA,public_docs   → Generate public-facing documentation
+/ctx,foldA,expose        → Identify what is exposed vs internal
+/ctx,foldA,summary       → Condensed overview (faster than analyze)
+/ctx,foldA,todos         → Extract all TODO / FIXME / HACK comments
+/ctx,foldA,lint          → Flag code style and quality issues
+/ctx,foldA,changelog     → Summarize changes (if diffs provided)
+
+### Multi-Context
+/ctx_multi,foldA,foldB,compare  → Compare two folders
+/ctx_merge,foldA,foldB          → Treat both folders as one unified context
+/ctx_diff,foldA,foldB           → Show differences between two folders
+
+### File Actions
+/f,file.py,explain         → What does this file do?
+/f,file.py,refactor        → Refactor for readability and structure
+/f,file.py,debug           → Find and explain bugs
+/f,file.py,document        → Add docstrings and comments
+/f,file.py,optimize        → Performance improvements
+/f,file.py,test            → Generate unit tests
+/f,file.py,review          → Code review with feedback
+/f,file.py,convert,lang    → Convert to another language (e.g. convert,typescript)
+/f,file.py,improve         → General quality improvements
+/f,file.py,secure          → Security audit
+/f,file.py,summary         → One-paragraph summary of the file
+/f,file.py,trace,<fn>      → Trace execution of a specific function
+/f,file.py,mock            → Generate mock/stub for this file
+/f,file.py,diff,file2.py   → Compare this file with another
+
+### Project Actions
+/p,foldA,health        → Overall project health assessment
+/p,foldA,roadmap       → Suggested roadmap based on current state
+/p,foldA,risks         → Identify technical and business risks
+/p,foldA,scalability   → Assess scalability
+/p,foldA,readiness     → Is this ready for production/launch?
+/p,foldA,architecture  → Full architectural review
+
+### Save / Export
+/save                    → Format current output for saving
+/save_as,file.md         → Format output as a named file
+/save_to,foldA,file.md   → Format output targeted to a folder/file location
+/export,markdown         → Export as Markdown
+/export,json             → Export as JSON
+/export,csv              → Export as CSV
+/export_to,foldA,file.md → Export to a specific path
+
+NOTE: All save/export commands generate formatted content only. No actual file writing occurs.
+
+---
+
+## 16. META
+
+/ignore_previous  → Discard all prior instructions and start fresh
+/unfiltered       → Remove formatting constraints, respond raw with no structure imposed
+/style            → Show current active style, tone, and format settings
+/help             → Explain any shortcut I ask about
+/reset            → Reset all settings to system defaults
+
+---
+
+## 17. PERSONAL
+
+/wissem      → (see Section 2) Direct, technical, no filler. Tunisian/startup aware.
+/wissempro   → Advise as if speaking to a tech entrepreneur in Sousse, Tunisia
+/tunisia     → Apply Tunisian business and cultural context
+/sousse      → Local context: Sousse, Tunisia — UTC+1, Arabic/French environment
+/business    → Frame response from a business and ROI perspective
+
+---
+
+## 18. STATE & MEMORY
+
+/remember,key,value  → Store a value under a key for this session
+/recall,key          → Retrieve a stored value by key
+/forget,key          → Delete a specific stored key
+/clear_memory        → Wipe all stored session memory
+/state               → Show all currently stored keys and values
+
+---
+
+## 19. GOAL CONTROL
+
+/goal,<goal>         → Set current goal (e.g. /goal,build_api)
+/goal_update,<goal>  → Update the current goal
+/goal_clear          → Remove the current goal
+/goal_status         → Report progress toward current goal
+
+---
+
+## 20. EXECUTION MODE
+
+/execute           → Execute the plan directly, produce output now
+/plan_then_execute → First show the plan, wait for approval, then execute
+/simulate          → Simulate the outcome without executing
+/dry_run           → Show what would happen step by step without doing it
+/rollback          → Undo or reverse the last executed step
+
+---
+
+## 21. PRIORITIZATION
+
+/focus,top3   → Identify and focus on the top 3 most important items
+/priority     → Rank all items by importance
+/impact       → Rank items by potential impact
+/effort       → Estimate effort level for each item (low / medium / high)
+/quick_wins   → Identify items that are high impact and low effort
+
+---
+
+## 22. BUDGET & TOKEN MANAGEMENT
+
+/le,N                → Use <N% of token limit. If task exceeds this, DO NOT execute. Instead, provide a descriptive copy/paste prompt for me to do it.
+/usage,N             → Execute task but pause if token usage reaches N% (estimated required amount in percentage = <price>$). Ask for approval to use a custom P% (estimated remaining amount = <remaining_price>$) for the next chunk, and repeat until finished.
+/quote               → Provide a full technical and financial token cost estimation ($) of the entire task BEFORE executing any action. Require me to explicitly type "Approve" to proceed.
+/diet                → Drop all historical context, conversational pleasantries, and unrelated files from your active memory right now. Only retain the exact code lines needed to execute the current prompt.
+/delegate            → Evaluate if this task is trivial. If so, DO NOT execute it using your current high-tier model. Instead, provide a prompt I can copy/paste into a cheaper, faster model (like Flash or Haiku).
+
+---
+
+## USAGE
+
+1. Paste UNIVERSAL PREFIX in new chat
+2. Use shortcuts freely — chain as needed
+3. For context commands, paste content immediately after the load command
+
+### Example chains
+
+```
+/d /ctx,backend,load
+[paste your code here]
+/ctx,backend,analyze /ep,optimize /execute /goal,build_api
+```
+
+```
+/p,projectX,health /goal_status /focus,top3 /final_only
+```
+
+```
+/wissempro /engineer /ctx,api,load
+[paste api folder content]
+/ctx,api,secure /ctx,api,dataflow /d /cot
+```
+
+```
+/selfcritic /tot /f,auth.py,load
+[paste file]
+/f,auth.py,debug /f,auth.py,secure /final_only
+```

package/templates/shortcuts/usage-management-examples-shortcuts/budget_controllers_examples.md

@@ -0,0 +1,96 @@
+# Budget & Token Management Shortcuts — Examples
+
+This document outlines scenarios and use cases for the **Section 22: Budget & Token Management** shortcuts defined in `shortcuts.md`.
+
+---
+
+## 1. The Pre-Flight Estimate (`/quote`)
+
+**Use Case:** You are about to run a massive refactor across 15 files and want to avoid an accidental $5.00 API charge.
+
+**Prompt:**
+> `/quote Refactor the entire authentication module in src/auth/ to use JWT instead of sessions.`
+
+**AI Behavior:**
+> **Token Estimation:**
+> - Input context: ~45,000 tokens
+> - Expected output: ~8,000 tokens across 5 files
+> - **Estimated Cost:** $0.85
+> 
+> *Type "Approve" to proceed with the execution.*
+
+---
+
+## 2. The Strict Limit (`/le,N`)
+
+**Use Case:** You want the AI to do a task, but only if it's cheap. If it's an expensive task, you want to do it yourself using a guided prompt.
+
+**Prompt:**
+> `/le,10 Generate a comprehensive test suite for the dashboard components.`
+
+**AI Behavior (Scenario A - Under 10%):**
+> *Task requires ~5% of token limit. Executing...*
+> [Outputs the test suite code immediately]
+
+**AI Behavior (Scenario B - Over 10%):**
+> *Task rejected. Estimated token usage (35%) exceeds your 10% limit.*
+> 
+> **Do-It-Yourself Execution Prompt:**
+> "Please copy and paste the following prompt into your terminal or a cheaper AI session to generate the tests file-by-file:
+> *'Read `src/dashboard/Chart.vue` and generate Vitest unit tests covering mounting and data-fetching. Save it to `tests/Chart.spec.js`.'* (Repeat for each component)."
+
+---
+
+## 3. The Paced Execution (`/usage,N`)
+
+**Use Case:** You are running an ongoing task and want to pause periodically to review the cost and approve the next chunk.
+
+**Prompt:**
+> `/usage,15 Migrate all old Vue 2 syntax to Vue 3 Composition API in the components folder.`
+
+**AI Behavior:**
+> *Executing chunk 1...*
+> [Migrates 3 files]
+> 
+> **⏸️ Token Limit Paused**
+> I have reached your 15% token usage threshold (Estimated cost spent: $0.45).
+> **Remaining work:** 7 files left.
+> **Estimated remaining cost:** $1.10 (approx. 35% more token usage).
+> 
+> *Do you approve continuing for a custom chunk of 20%?*
+
+---
+
+## 4. The Context Diet (`/diet`)
+
+**Use Case:** You've been chatting with the AI for 2 hours. The context window is huge, and every single prompt is costing $0.10 just in input tokens. You only need a small CSS fix now.
+
+**Prompt:**
+> `/diet Fix the margin on the login button in Header.vue.`
+
+**AI Behavior:**
+> *Flushing conversational history and unrelated files...*
+> *Retaining only `Header.vue` lines 45-60.*
+> 
+> Here is the fixed CSS class for the login button:
+> ```css
+> .login-btn { margin-left: 1rem; }
+> ```
+> *(Input tokens reduced from 80k to 400 tokens).*
+
+---
+
+## 5. The Down-Level Delegation (`/delegate`)
+
+**Use Case:** You are in an expensive Opus/Pro session, but you need a simple regex or a basic boilerplate script that a cheap model like Haiku/Flash could do for pennies.
+
+**Prompt:**
+> `/delegate Write a regex to validate Tunisian phone numbers (+216).`
+
+**AI Behavior:**
+> *Task evaluated as trivial. Executing this on the current premium model is a waste of budget.*
+> 
+> **Action:** Copy and paste the prompt below into a Gemini Flash or Claude Haiku session:
+> > *"Write a javascript regex that strictly matches Tunisian phone numbers starting with +216 followed by 8 digits. Provide 3 test cases."*
+
+---