UNPKG

major-ai-skills

Version:

Installable agentic skills / AI agent skills (SKILL.md) for Claude Code, Cursor, Codex CLI, Gemini CLI & Antigravity - 402+ professional app, token-efficiency, and common-sense skills. SEO/GEO ready.

99 lines (72 loc) 4.64 kB
--- name: flag-missing-facts-first description: "Identify facts that are required for a reliable answer and distinguish them from optional context." category: common-sense risk: safe source: self source_type: self date_added: "2026-08-26" tags: ["fact-checking", "information-readiness", "clarification", "pre-flight-check", "debugging", "prompt-engineering"] tools: ["claude", "cursor", "gemini", "codex", "chatgpt"] --- # Flag Missing Facts First (The 10/10 Information Readiness Check) (AI Skill) ## Overview When users provide incomplete data to an AI assistant (*"My Python script crashes with an error, how do I fix it?"*), most models begin guessing wild possibilities rather than asking for the stack trace or environment details. The **10/10 Information Readiness Check** instructs the AI to pause, identify the exact missing variables required to deliver a perfect, definitive answer, and request them in a structured intake checklist. --- ## The Information Readiness Gate ``` ┌─────────────────────────────────────────────────────────────┐ Information Readiness Gate User Initial Prompt (Incomplete Data) [ PRE-FLIGHT AUDIT: Rate Information Completeness /10 ] [ List 3-4 Missing Concrete Facts Needed for a 10/10 ] User Supplies Missing Facts ──► 100% Definitive Solution └─────────────────────────────────────────────────────────────┘ ``` --- ## Master Information-Audit Prompt Templates ### Pattern 1: The "10/10 Readiness" Directive Use before launching any complex analysis, debugging, or proposal: ```markdown I want to [INSERT PROJECT / TASK: e.g. Design our database sharding strategy]. Here is what I have so far: [PASTE CURRENT NOTES / DATA]. Do NOT attempt to write the final solution yet. 1. Rate the completeness of my prompt on a scale of 1 to 10. 2. List the **top 3 missing facts or parameters** you need from me to give a 10/10 production-ready response. 3. For each missing fact, give an example of what a good answer looks like. Wait for my response. ``` --- ### Pattern 2: The Bug & Diagnostic Intake Form Use for engineering, IT support, or system troubleshooting: ```markdown I have a bug in my application: [BRIEF SUMMARY]. Act as a Principal Support Engineer: - Generate a 4-point diagnostic intake checklist of the exact error logs, environment variables, and reproduction steps you need to pinpoint the root cause. ``` --- ## Real-World Case Study ### Scenario: Troubleshooting a Slow React Application #### Weak Prompt (Guesswork & Speculation) > **User**: *"My React app is rendering slowly when users scroll the feed. How do I fix it?"* > > *AI guesses 10 generic optimizations: useMemo, useCallback, React.memo, virtualized lists, image lazy loading, debounce scroll, code splitting... (Leaves the user overwhelmed with trial-and-error).* #### 10/10 Readiness Prompt & Structured Intake > **User**: *"Before suggesting fixes, what 3 profiling facts do you need to diagnose why my React feed is lagging?"* **AI Response**: > To give you the exact fix rather than generic advice, please provide: > 1. **Component Tree Depth & State Location**: Is the scroll state held in the root parent or localized? > 2. **DOM Node Count**: How many DOM nodes are rendered simultaneously when scrolling (e.g. 50 items vs 5,000 items)? > 3. **React DevTools Profiler Output**: Which specific component shows the highest "Render duration" flamegraph bar? **Result**: The user checked the Profiler, found that a single un-memoized SVG icon component was re-rendering 2,000 times, and fixed it in 30 seconds. --- ## The 10/10 Rule > **"1 minute spent providing the missing facts saves 30 minutes of debugging generic AI hallucinations."**