# AI Productivity

> AI Productivity

- Skill: `comcclelland/ai-productivity` (Agent Skill)
- Install (CLI): `npx skillmds@latest add comcclelland/ai-productivity`
- Raw SKILL.md: https://api.skillmd.com/api/skills/comcclelland/ai-productivity/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: comcclelland (https://skillmd.com/u/comcclelland)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/comcclelland/ai-productivity

---

# AI Productivity

Structures vague prompts into concrete, actionable tasks. Compresses context for efficient execution.

## When to Use

- When the user gives a vague or broad request ("make this better", "clean this up", "I need help with...")
- When a task has many possible interpretations
- When context is scattered and needs organizing before action

## Process

### Step 1: Clarify Intent
- Identify what the user actually wants (outcome, not activity)
- Ask ONE clarifying question if intent is genuinely ambiguous
- If intent is clear enough to start, don't ask — just propose

### Step 2: Decompose
Break the request into concrete, ordered tasks:
- Each task should be completable in one action
- Tasks should be independent where possible
- Identify dependencies (what must happen first)
- Estimate relative effort (quick / medium / involved)

### Step 3: Prioritize
Order tasks by:
1. Blockers first (unblocks other work)
2. High-impact, low-effort next (quick wins)
3. High-impact, high-effort last (planned work)
4. Low-impact → defer or skip

### Step 4: Execute
- Start with the first actionable task
- Report progress after each task completes
- Adjust the plan if new information emerges
- Don't over-plan — act, learn, adapt

### Rules
- Bias toward action over analysis
- Never ask more than one question before starting
- If you can reasonably infer intent, do so and state your assumption
- Keep task descriptions short and imperative ("Add validation to X", not "We should consider adding...")
- If a task is too large, split it — if too small, combine it