- Home
- Skills
- Testing & Debugging
- prod-bug-hunter
Production Bug Hunter
Systematic runtime bug detection and automated patching for Firebase, Next.js, and Vercel stacks.
Free
Works with the AI Agents you already use
prod-bug-hunter
Example session with this skill installed
Run a priority P2 sweep on my hooks directory to find any Firestore listener leaks or missing cleanups.
- Read your context and instructions
- Compiled the prod-bug-hunter
SEVERITY: P2
FILE: src/hooks/useChat.ts:42
CAUSE: onSnapshot listener lacks cleanup, causing memory leaks on unmount.
FIX
useEffect(() => {
const unsub = onSnapshot(docRef, (doc) => setData(doc.data()));
return () => unsub(); // Added cleanup
}, [docId]);
Connects securely to your tools. The creator never sees your data.
About this skill
Stop Production Fires Before They Start
The Production Bug Hunter is a specialized diagnostic skill designed to bridge the gap between "tests pass" and "production works." It targets the complex runtime failures—race conditions, stale states, and listener leaks—that static analysis and unit tests often miss in high-load environments.
What it does
This skill performs a priority-ordered sweep of your repository to identify and fix critical infrastructure bugs tailored to the Firebase, Next.js, and Vercel stack. It specifically hunts for:
- P0 Auth Race Conditions: Detects simultaneous Firebase initialization and unguarded auth state access.
- P1 Stale State Drift: Identifies UI/database desyncs and missing optimistic update rollbacks.
- P2 Listener Leaks: Finds uncleaned
onSnapshotsubscriptions that cause memory bloat and stale data. - P3 Environment Mismatches: Cross-references source code with CI/CD YAML files to catch missing secrets and version drifts.
Why use this skill?
Unlike general-purpose coding assistants, this skill employs a systematic "hunter-killer" methodology. It doesn't just look for clean code; it looks for architectural patterns that break under real-world stress. For a developer, it's like having a senior SRE perform a deep-dive audit of your app's reactivity and auth patterns in seconds, providing ready-to-apply patches for every vulnerability found.
- Requires manual verification of file paths.
- Cannot trigger live deployment tasks.
- Limited to scanning 50k tokens per sweep.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
This skill has been used 20 times.
Trust & safety
Security scan
Security scan passed 7 months ago
- Free with an Agensi account
Requires
Permissions
Allowed Hosts