Agent Assistant — Debug (Hard)

What it does

Provides a full ‘/debug:hard’ workflow for the Agent Assistant project. When triggered it instructs the agent to load core rules, spawn the configured multi-agent environment, and follow the repository's debug command scripts to deeply investigate complex problems. The skill centralises policy (CORE.md), agent orchestration, and skill resolution so the debugging run is consistent and repeatable.

When to use it

Use this skill when a user explicitly requests /debug:hard or when an issue requires a deep, multi-agent investigation (complex bugs, production incidents, hard-to-reproduce failures). It is intended for cases where lighter debugging steps have failed and an end-to-end diagnostic run is needed.

What's included

• Scripts: none detected in the skill directory, but the workflow references repository agents and command files under the antigravity project.
• References: no dedicated references folder in the skill, but it points to CORE rules and command markdown under the repo.
• Instructions: the SKILL.md instructs loading the CORE rules, spawning Codex-native multi-agents, and resolving matrix skills before running the debug command.

Compatible agents

Inferred compatibility: Codex / Gemini-style multi-agent runtimes and agents that support repository-based skill resolution (Codex, Gemini, similar agent frameworks).