A gamified leaderboard for Kiro IDE users — track your prompts, tool calls, credits, and session time. Climb the ranks. Earn badges. Roast your friends.
You and your team write code with Kiro. You ship features. You crush bugs. But nobody knows who's actually pulling their weight in the AI revolution.
Kironomics fixes that.
It's a private leaderboard that quietly tracks how much you actually use Kiro — the prompts you fire off, the tools the AI calls, the credits you burn through, the sessions that go past midnight. Then it ranks everyone, hands out badges, and shows who's the real Session Immortal of the team.
Think Strava for AI coding. Or GitHub contributions, but spicier.
| 🥇 Compete with friends | Daily, weekly, all-time leaderboards |
| 🏅 Earn 7 badges | First Session, Century Club, Marathon Runner, and more |
| 👑 Unlock 7 titles | Climb from "Kiro Novice" 🌱 to "Session Immortal" 🏆 |
| 🔥 Streak tracking | GitHub-style heatmap of your last 365 days |
| 💳 Real credit tracking | See your Kiro plan usage live (Free/Pro/Pro+/Power) |
| 📊 Burn rate analytics | "You'll run out of credits in 9 days at this pace" |
| 🛡️ Anti-cheat verified | Reads your actual AWS-synced credit data |
| 🤫 Zero overhead | Everything batches at session end |
In Kiro, open the Powers panel (left sidebar) and click + Add Custom Power.
Option A — From GitHub (recommended):
https://github.com/logesh4v/Kironomics
Option B — From local folder: Point to the path of this repo on your machine.
Open the Kironomics dashboard → Settings → click Save Changes → copy the API key shown.
In Kiro chat, just say:
"Set up Kironomics"
Kiro will read POWER.md, ask you for your API key + backend URL, and create the 4 tracking files in your workspace automatically.
Tracking is silent and passive. When the agent stops, one batched request goes to the leaderboard.
Open the dashboard whenever you want. Watch yourself climb.
This is a Knowledge Base Power — pure documentation. When you activate it, Kiro reads the POWER.md instructions and creates these 4 files in your workspace:
.kiro/
├── kironomics_report.py (Python session reporter)
└── hooks/
├── kironomics-tool-counter.kiro.hook (counts every AI tool call)
├── kironomics-prompt-counter.kiro.hook (counts every prompt + records start time)
└── kironomics-session-reporter.kiro.hook (sends data on agentStop)
Then on every session:
┌────────────────────────────────────────────────────────────────────┐
│ YOUR LAPTOP │
│ │
│ ┌──────────────┐ auto-writes ┌────────────────────────┐ │
│ │ Kiro IDE │ ─────────────────►│ state.vscdb │ │
│ │ │ │ └─ kiro.kiroAgent │ │
│ │ │ │ └─ usageState │ │
│ └──────────────┘ │ ├─ currentUsage│ │
│ │ │ ├─ usageLimit │ │
│ │ │ └─ resetDate │ │
│ ▼ └────────────────────────┘ │
│ ┌────────────────────────────────────────┴────────────────┐ │
│ │ 3 Silent Hooks │ │
│ │ 1️⃣ Tool Counter → /tmp/kironomics_tools │ │
│ │ 2️⃣ Prompt Counter → /tmp/kironomics_prompts │ │
│ │ 3️⃣ Session Reporter (agentStop): │ │
│ │ runs Python script: │ │
│ │ • reads /tmp counters │ │
│ │ • reads state.vscdb credit data │ │
│ │ • POSTs to backend │ │
│ └─────────────────────────────────────────────────────────┘ │
└──────────────────────────┬──────────────────────────────────────┘
│ HTTPS POST (1 request per session)
▼
┌──────────────────────────────────────────────────────────────────┐
│ 📊 KIRONOMICS BACKEND │
│ • Validates token │
│ • Detects plan (Free / Pro / Pro+ / Power) │
│ • Calculates score │
│ • Updates leaderboard │
└──────────────────────────────────────────────────────────────────┘
| Title | Score Range | Icon |
|---|---|---|
| Kiro Novice | 0 – 99 | 🌱 |
| Prompt Apprentice | 100 – 499 | 📝 |
| Hook Wrangler | 500 – 1,999 | 🪝 |
| Tool Wielder | 2,000 – 4,999 | 🔧 |
| Session Warrior | 5,000 – 14,999 | ⚔️ |
| Token Legend | 15,000 – 49,999 | 👑 |
| Session Immortal | 50,000+ | 🏆 |
Score formula (credit-based, the cheat-proof one):
score = (credits_consumed × 100) + (tool_calls × 1) + (prompts × 2)
Fallback for users without credit data yet:
score = (tool_calls × 3) + (prompts × 5) + (elapsed_minutes)
| Badge | Criteria | Icon |
|---|---|---|
| First Session | 1 session completed | 🎯 |
| Century Club | 100 tool calls | 💯 |
| Hook Hero | 50 prompts | 🦸 |
| Token Titan | 500 tool calls | 💎 |
| Marathon Runner | 10+ hours total | 🏃 |
| Week Warrior | 7-day streak | 🔥 |
| Monthly Master | 30-day streak | 🌟 |
Kironomics reads your actual Kiro usage straight from Kiro's local database (state.vscdb) — the same data Kiro shows in its own UI.
That means:
- 📊 Live credit balance — see how many credits you've burned this month
- 📈 Burn rate — "73 credits/day at current pace"
- ⏰ Days remaining — "9 days until your monthly reset"
- 🎯 Plan auto-detected — Free, Pro, Pro+, Power, Auto-Auto
- 🛡️ Tamper-proof — you can't fake AWS billing data
Personal Metrics card preview:
┌────────────────────────────────────────────────────┐
│ 📊 Plan & Credits Kiro Pro+ $40/mo │
│ Live data from Kiro IDE │
│ │
│ 1,246.79 / 2,000 credits 62.3% │
│ ████████████░░░░░░░░ │
│ │
│ ┌────────┬──────────┬──────────┬──────────────┐ │
│ │ 753 │ 17 days │ 73/day │ 9 days │ │
│ │ Remain │ Resets in│ Burn rate│ Will run out │ │
│ └────────┴──────────┴──────────┴──────────────┘ │
└────────────────────────────────────────────────────┘
The credit data comes from Kiro's own SQLite database, server-synced from AWS. Cheating attempts:
- Fake hook counters? → Backend cross-references with
state.vscdbcredit deltas. Caught. - Hooks claim 1000 events but credits show 0? → Sanity check fails after 3 zero-credit sessions. Flagged.
- Credits-to-events ratio over 50x? → Impossible even on heavy Opus. Flagged.
- Optional AWS Cost Explorer verification → Cross-reference with actual AWS billing for ✅ Verified badge.
Cheaters get hidden from the leaderboard. The honest crew rules.
Q: Will this slow down Kiro?
No. The 3 hooks fire in <2ms each (just incrementing a number in a temp file). The actual data send happens once when the agent stops — not during your work.
Q: What does it actually track?
- Number of prompts you submit
- Number of tools the AI calls (read_file, str_replace, execute_bash, etc.)
- Session duration
- Credits consumed (real, from Kiro's own data)
- Your plan (Free / Pro / Pro+ / Power)
Nothing about what you typed or what the code looked like. Just counts.
Q: Where does the credit data come from?
From state.vscdb — Kiro's own SQLite database that lives at:
~/Library/Application Support/Kiro/User/globalStorage/state.vscdb
We open it in read-only mode, so Kiro IDE is never affected.
Q: I'm on the Free plan. Will I see plan info?
Yes — Kironomics auto-detects all plans (50 credits = Free, 1000 = Pro, 2000 = Pro+, 10000 = Power). Auto-Auto users skip credit tracking.
Q: Do I need to share AWS credentials?
No. Credit data is read locally from your machine. AWS Cost Explorer verification is optional and only used for the ✅ AWS Verified badge.
Q: Can I uninstall it?
Yes. Just delete .kiro/hooks/kironomics-* and .kiro/kironomics_report.py. Or remove the Power from Kiro.
Q: Why no MCP server?
This is a Knowledge Base Power following Kiro's official pattern. POWER.md tells Kiro the agent exactly what to create, and Kiro does it. Simpler, more reliable, and works in any workspace.
.
├── POWER.md ← Power manifest (Kiro reads this on activation)
└── README.md ← This file (for GitHub viewers)
That's it. Two files. No MCP server, no node_modules, no path issues.
When you activate the Power, Kiro reads POWER.md and creates the tracking files in your workspace.
The Kironomics backend and dashboard live in a separate repo. To run them:
Backend (Express.js):
cd backend
npm install
npm start
# → http://localhost:8080Frontend (React + Vite):
cd frontend
npm install
npm run dev
# → http://localhost:3000Environment:
# Frontend
VITE_API_BASE_URL=http://localhost:8080
VITE_COGNITO_USER_POOL_ID=us-west-2_XXX
VITE_COGNITO_CLIENT_ID=XXXPRs welcome. Ideas worth submitting:
- New badges
- Team challenges
- Slack/Discord integration
- More dashboard charts
- A roast feature for the bottom of the leaderboard 😈
MIT — go wild.
Built with ❤️ and probably too many late-night Kiro sessions.
⭐ Star this repo if you'd actually use it.