s-debug_skill

Overview

This skill diagnoses and fixes World of Warcraft addon bugs using evidence-driven investigation. It enforces runtime proof before proposing changes, avoiding guesses from static code alone. The skill covers Lua errors, taint and combat lockdown issues, instrumentation logging, and API failures common around major client updates.

How this skill works

It guides a hypothesis-driven workflow: gather runtime evidence, isolate the fault, form a testable hypothesis, instrument with logs, and validate fixes with repeatable reloads. It uses MCP tools for capturing output, queuing Lua evals, and retrieving results, and prefers MechanicLib:Log for shareable, filterable debug output. The skill interprets stack traces, taint paths, and protected-frame failures to identify root causes.

When to use it

• You see runtime errors like "attempt to index nil value" or nil dereferences.
• When addons trigger "Action blocked by Blizzard" or other taint-related behavior.
• If features fail only in combat or protected frames display bugs.
• After a crash or repeatable error when you need reliable evidence before editing code.
• When API deprecations or changes around major patches may be the root cause.

Best practices

• Always collect runtime output via addon.output(agent_mode=true) after a /reload before changing code.
• Use MechanicLib:Log() for instrumentation so logs are copyable and accessible to the agent.
• Form concise hypotheses (If X then Y because Z) and design a single test to falsify each hypothesis.
• Isolate the addon: reproduce with minimal code and disabled dependencies to reduce noise.
• Use MCP tools (lua.queue, lua.results, addon.lint) rather than ad-hoc shell commands for reproducible evidence.

Example use cases

• Diagnose a nil-value crash by capturing addon.output and inspecting the stack trace to find the missing API result.
• Track down taint propagation after an "Interface action failed" report by instrumenting frame creation and secure hooks.
• Resolve combat lockdown failures by identifying protected calls attempted during combat and moving them to secure templates or event-driven handlers.
• Verify an API change after a patch by queuing Lua evals to check return values and behavior across units.