The build-decision guide for the AI era

Should it be a
prompt?

Or a skill, an MCP server, an agent, a real app, or nothing at all. Answer eight questions and get a verdict with the reasoning, the trade-offs and a first step.

By Mahmoud Halat, founder of Space & Story. Free, no sign-up, no model call: the logic is published.

  1. 6App / systemWeeks to months
  2. 5AgentDays to weeks
  3. 4MCP / connectorHours to days
  4. 3SkillAn afternoon
  5. 2Saved promptUnder an hour
  6. 1PromptMinutes
× Don't build it often the right answer
Start at the bottom. Climb only when a question forces you to.

Step one

Run your task up the ladder

Eight questions. The cheapest rung that works wins, and every verdict shows the rule that set it. How the rules work.

Loading the tool…

What the patterns say

15 of 38 everyday tasks are skills. Only 7 need an agent or an app.

We ran 38 common work tasks, from meeting notes to invoice matching, through the same rules as the tool. Most land low on the ladder: packaged instructions beat software far more often than the hype suggests. Each one has its reasoning and a starting point.

Open the pattern library
Number of the 38 patterns that land on each rung
  1. Prompt3Alt text for a site's image backlog · Explaining one clause you don't understand
  2. Saved prompt3Weekly status update to your manager · Social posts from a new blog article
  3. Skill15Meeting transcript to summary and action items · Proposal and quote drafting
  4. MCP / connector7Weekly KPI report from your analytics · Sales call prep brief
  5. Agent4Inbound lead triage and routing · Churn-risk alerts for account managers
  6. App / system3Three-way invoice matching (PO, receipt, invoice) · Personalised cold outreach at scale
  7. Don't build3Categorising expenses and bank transactions · Drafting replies in your own inbox

The explainer

Prompt vs skill vs MCP server vs agent vs app: what each one is, what it costs, when it breaks.

Read the definitions
Effort and upkeep for each rung
RungFirst versionUpkeep
PromptMinutesNone
Saved promptUnder an hourEdit when your process changes
SkillAn afternoonAn owner, a version, a test case or two
MCP / connectorHours to daysAuth, permissions, API changes
AgentDays to weeksEvals, logs, cost limits, someone on call
App / systemWeeks to monthsA team or a vendor, forever

Comparisons

Comparisons, from published specs

Prices and features checked against official pages in October 2026, with links.

  1. 01Claude Skills vs custom GPTs vs Gemini Gems vs Copilot agentsWhat Claude Skills, custom GPTs, Gemini Gems (now Gemini skills) and Microsoft 365 Copilot agents can do, who can build and share them, and the plan you need. Checked Oct 2026.6 min readChecked Oct 31, 2026
  2. 02AI app builders compared: Lovable, Bolt, v0, Replit Agent, Claude Code and CursorSix AI app builders compared on pricing (checked Oct 2026), strengths, weak spots and who should pick which, plus the signs you've outgrown them.7 min readChecked Oct 31, 2026
  3. 03Where to find Claude skills and MCP serversAnthropic's skills repo, Claude's directory, skills.sh, Skillspace, the official MCP Registry, Glama, Smithery and mcp.so compared on size, quality signals and safety.6 min readChecked Oct 31, 2026

Essays

The Moat Line essays

The other half of the question: once you can build it in a weekend, can anyone else?