silent-failure-hunter

What This Skill Does

The silent-failure-hunter is an advanced code auditing agent designed to identify, analyze, and remediate error handling anti-patterns within your codebase. In production environments, silent failures represent a significant technical debt that leads to data inconsistencies, corrupted states, and frustrated users. This skill automates the reconnaissance of your try-catch blocks and promise catch handlers, systematically enforcing five rigorous safety rules. It moves beyond simple syntax checking by evaluating whether your error handling strategies provide sufficient visibility to developers and appropriate context to the end-user. By detecting swallowed errors, console-only logging, and dangerous silent fallbacks, this agent ensures your application's resilience is robust and observable.

Installation

To integrate the silent-failure-hunter into your OpenClaw workflow, execute the following command in your terminal within your project root:

clawhub install openclaw/skills/skills/anton-abyzov/sw-silent-failure-hunter

Ensure that you have sufficient permissions to allow the agent to scan your source directory.

Use Cases

• Production Audits: Use this agent before major releases to identify hidden failure points that could cause outages.
• Legacy Code Refactoring: When touching legacy modules, use this tool to uncover "buried" errors that have been masked by empty catch blocks for years.
• PR Reviews: Run this agent against feature branches to prevent the introduction of new, non-actionable error handling code.
• Incident Investigation: If users report inconsistent data, use this tool to hunt for catch blocks that might be returning null or stale data silently.

Example Prompts

1. "Analyze the authentication service for silent failures and suggest improvements for the error handling in our catch blocks."
2. "I suspect we have some swallowed errors in the payment processing module. Can you run the silent-failure-hunter on /src/services/payments and report any high-priority issues?"
3. "Review the recently modified database controllers. Ensure that no catch blocks are masking potential network timeouts or constraint violations."

Tips & Limitations

• Context is King: The agent performs best when it has access to the full repository. Ensure file-read permissions are granted.
• False Positives: Sometimes a silent catch is intentional (e.g., ignoring a specific, non-critical file system error). You may need to instruct the agent to ignore specific files or patterns after its initial sweep.
• Manual Oversight: While the agent is excellent at finding where problems exist, the fix often requires understanding business logic. Always review the agent's proposed remediation code before applying it to your codebase.