fix-bug

Diagnose silent integration failures with runtime-contract-first evidence and ranked hypotheses.

121|2|Updated Apr 24, 2026
One-click install
npx skills add https://github.com/AgentSystemLabs/core --skill fix-bug-agentsystemlabs
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fix-bug
Source: https://github.com/AgentSystemLabs/core/tree/main/plugins/agentsystem-core/skills/fix-bug
Command: npx skills add https://github.com/AgentSystemLabs/core --skill fix-bug-agentsystemlabs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

fix-bug diagnoses integration and feature failures where the system runs but the expected side effect never occurs, such as hooks not firing, status not updating, callbacks not arriving, jobs not running, or webhooks going silent.

Core Features & Use Cases

  • Runtime contract-first tracing: maps the end-to-end trigger → dispatch → receiver → observable state using file:line evidence before forming hypotheses.
  • Silent-failure detection: searches for swallow patterns like || true, empty catch blocks, and redirected stderr/stdout that hide the real error.
  • Regression-focused mode: when something used to work, switches to a git-history driven workflow (log/blame/bisect) using the provided “worked yesterday / broken since …” anchor.

Quick Start

Tell the AI that the expected hook/status/webhook “didn’t trigger” and include any version anchor, then ask it to run fix-bug with mode balanced (or mode regression if it used to work).

Frequently Asked Questions about fix-bug

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I debug webhook callbacks that fail silently without throwing any errors?

Debug silent webhook failures by mapping the end-to-end runtime contract from trigger to observable state, identifying swallow patterns like empty catch blocks or redirected stderr, and running a single disambiguating diagnostic to isolate the hidden error.

Why does my cron job or queue worker not trigger the expected side effect despite no errors?

Cron misfires and queue worker stalls often stem from swallowed exceptions or env-var injection issues; trace the dispatch path using file:line evidence to find where the runtime contract breaks before forming ranked hypotheses.

What's the best way to find a silent integration failure that used to work but recently broke?

Find silent integration regressions by switching to a git-history driven workflow using log, blame, and bisect anchored on a version where the feature last worked, then trace the runtime contract to pinpoint the introduced swallow pattern.

How do I detect hidden errors in config file writes or env-var injection issues?

Detect hidden config write and env-var injection errors by mapping the expected runtime contract of endpoints and artifacts, then searching for silent-failure sites like `|| true` operators, empty catch blocks, or redirected stdout that mask the real error.

Does this debugging approach work for file-watcher triggers and tool-call delivery problems?

Yes, the runtime-contract-first method applies to file-watcher triggers and tool-call or MCP delivery problems by requiring explicit identification of silent-failure sites and a single diagnostic to disambiguate whether the dispatch or receiver stage dropped the event.

What should I do after diagnosing a silent failure to prevent it from happening again?

After diagnosing a silent failure, apply the ranked fix with the highest confidence label and perform optional regression test hardening to ensure the swallow pattern or missing side effect triggers a visible error in future runs.