using-roach

Enforces skill invocation before any response in Claude Code sessions.

1|Updated Feb 10, 2026
One-click install
npx skills add https://github.com/stefanfaur/roach-marketplace --skill using-roach-stefanfaur
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: using-roach
Source: https://github.com/stefanfaur/roach-marketplace/tree/main/roach/skills/using-roach
Command: npx skills add https://github.com/stefanfaur/roach-marketplace --skill using-roach-stefanfaur

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI assistants often skip relevant workflow skills and answer directly, leading to inconsistent processes and missed best practices. This Skill establishes a mandatory rule that relevant skills must be invoked before any response, including clarifying questions or codebase exploration. ## Core Features & Use Cases - Mandatory Skill-First Rule: Requires invoking any potentially relevant skill via the Skill tool before responding, even for simple questions. - Skill Priority Ordering: Defines that process skills (brainstorming, systematic-debugging) run before implementation skills when multiple apply. - Unified Workflow Map: Organizes skills into research/design, execution, quality, and continuity phases with domain-scoped artifact storage under thoughts/shared/. - Use Case: When a user says "fix this bug", the assistant invokes systematic-debugging first rather than jumping straight into editing code. ## Quick Start Start any conversation and the assistant will check for and invoke applicable skills before responding to your request.

Frequently Asked Questions about using-roach

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

FAQPage Schema
How do I make Claude Code always check skills before answering?▼

Install the roach plugin and its using-roach skill, which mandates invoking any relevant skill via the Skill tool before any response. This applies even to clarifying questions and file exploration, with a 1% relevance threshold triggering invocation.

What is the correct skill order for building a feature?▼

Process skills come first: invoke brainstorming before implementation skills when building something new. For bug fixes, systematic-debugging runs first, followed by domain-specific implementation skills.

Does this skill apply to subagents dispatched for tasks?▼

No. The skill includes a SUBAGENT-STOP directive instructing subagents dispatched to execute specific tasks to ignore it. It governs only top-level conversation behavior.

Where are plans and research documents stored?▼

Artifacts are stored under thoughts/shared/ with domain-scoped paths: plans in plans/<domain>/, research in research/<domain>/, and handoffs in handoffs/<domain>/, using date-prefixed filenames. The domain is auto-detected from task context.

Why should I avoid EnterPlanMode and ExitPlanMode tools?▼

These tools trap the session in plan mode where Write and Edit tools are restricted. The brainstorming and writing-plans skills provide their own structured planning flow without this limitation.