A structured decision framework for evaluating shiny objects — new tools, platforms, AI developments, and opportunities that compete for your attention.
Built for founders, operators, and builders who run multiple ventures and live in a world where something new launches every day.
You see something exciting. Your brain lights up. You want to adopt it right now. But you're already running three things, your roadmap is full, and last month's "game changer" is sitting unused.
FOMO isn't a character flaw — it's a signal that you care about staying competitive. But without a filter, every signal gets the same weight, and your focus fractures.
The FOMO Analyzer runs shiny objects through a structured filter in two modes:
Five diagnostic questions that cut through excitement:
- Which venture does this serve? — Maps to your actual businesses
- What does this replace? — Additive load vs. genuine upgrade
- What's the cost of waiting 30 days? — Urgency check
- Who executes this? — If it's you, it competes with everything
- Does this move a current priority forward? — Roadmap alignment
Each answer scores green/yellow/red. The verdict is instant:
| Score | Verdict | What Happens |
|---|---|---|
| 4-5 greens | GO | Add to roadmap with a slot |
| 2-3 greens | PARK | Log it, revisit in 30 days |
| 0-1 greens | FOMO | Capture the idea, move on |
For anything that passes the quick filter, three purpose-built agents stress-test the decision:
| Agent | Lens | Question |
|---|---|---|
| The Realist | Execution cost | What does this ACTUALLY cost to adopt? |
| The Skeptic | Hidden risks | What breaks? What's the vendor risk? |
| The Mirror | Frame validity | Is this a real need, or a solution looking for a problem? |
Deep mode can override quick mode — a quick GO can become a PARK after the agents expose hidden costs.
Deep mode adds a fourth verdict: KILL — reserved for things that would actively harm your current plan.
Even FOMO verdicts carry a signal. After every evaluation, the analyzer asks: "What about this excited you — not the tool, but the capability?"
Then it checks your existing stack and suggests improvements at three levels:
| Level | Effort | Example |
|---|---|---|
| Quick Win | <1 hour | Enable an existing feature you haven't tried |
| Small Build | <1 day | New automation connecting two tools you already have |
| Project | >1 day | Meaningful build, but using your current stack |
This turns FOMO into a spark for building — the shiny object might be wrong, but the need it triggered might be real and addressable with what you already have.
Copy SKILL.md to your Claude Code skills directory:
mkdir -p ~/.claude/skills/fomo
cp SKILL.md ~/.claude/skills/fomo/SKILL.mdThen optionally create a config:
cp fomoconfig.example.yml ~/.claude/skills/fomo/fomoconfig.yml
# Edit with your ventures, tools, and file pathsCopy SKILL.md to your .github/ directory or instructions path as configured in your Copilot setup.
The SKILL.md file is a self-contained prompt. Feed it to any AI assistant that supports tool use and agent spawning. The five diagnostic questions work even without agents — just run them manually.
/fomo "Cursor looks amazing, should we switch from VS Code?"
/fomo --deep "This new CRM claims to replace our entire stack"
/fomo --log
/fomo --retro
The analyzer also watches for FOMO signals mid-conversation and offers a gentle nudge (once, never nagging).
Every evaluation is logged. Over time, the log reveals patterns:
- Which ventures trigger the most FOMO?
- Is it usually novelty, comparison anxiety, or genuine need?
- How many parked items actually get picked up?
- How many GO verdicts delivered real value?
Run /fomo --retro after 10+ entries for an honest retrospective.
Copy fomoconfig.example.yml to fomoconfig.yml and customize:
- ventures — your businesses/projects with priority levels
- tools — your current stack (checked for lane conflicts)
- fomo_log — where to store evaluation history
- parking_lot — where to park ideas for later
- priorities_source — your roadmap or task list for alignment checks
The analyzer works without a config (infers from conversation), but the config makes it faster.
- Works with impulsive energy, not against it — capture first, filter second
- Fast by default — quick mode takes 2 minutes, deep mode only when earned
- No idea is lost — everything gets logged, even FOMO verdicts
- Not a killjoy — some shiny objects ARE strategic; the filter finds them
- One nudge, max — never nags, never shames
- Patterns over incidents — the retrospective reveals your FOMO fingerprint
MIT