scout · onewave-ai — Claude Code skill: what it does, how to install
scout reads the running conversation and names the skill or subagent that fits the task, with a short plan for using them together.
MIT 314 installs ★ 440 updated 2026-10-20
npx skills add OneWave-AI/claude-skills --skill scout -a claude-codeWhat scout does
The skill turns the current thread into a recommendation. It looks at what you are doing, sorts the task into a type, then points at a skill to invoke or a subagent to hand the work to.
Its instructions are a prompt with a fixed answer shape: the task as read, the primary tool and the reason for it, optional helpers, and a numbered workflow. The list it draws on covers subagents such as Explore, Plan, Code Reviewer, Security Auditor and Compliance Auditor, plus slash commands grouped by area: development, documentation, analysis, marketing, creative. File reads and web lookups go through Read, Glob, Grep, WebFetch and WebSearch; the model follows whatever the session uses.
The subagent names and commands in that list are Claude Code ones, Task tool included. Run it under Cursor or Codex and it will still name them, so expect suggestions you cannot invoke as written.
What it helps with
- pick the right skill when a repo ships dozens of them
- hand a refactor to Code Reviewer instead of writing the prompt yourself
- lay out a multi-step workflow across several subagents
- recall the exact slash command for scaffolding an API endpoint
- split a security pass between Security Auditor and Dependency Auditor
Example prompts for the agent
/scout— which of these skills fits a database migration?- Can't tell whether to use a subagent or a slash command here. Recommend one and say why.
- Break this task into steps and name the tool for each.
How to install scout
1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill scout -a claude-code2Manually: put the files in place
~/.claude/skills/scout/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/scout/
1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill scout -a codex2Manually: put the files in place
.agents/skills/scout/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill scout -a cursor2Manually: put the files in place
.cursor/skills/scout/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install OneWave-AI/claude-skills scout2Manually: put the files in place
.github/skills/scout/SKILL.md1Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill scout -a gemini-cli2Platform CLI
gemini skills install https://github.com/OneWave-AI/claude-skills --path scoutManually: put the files in place
.devin/skills/scout/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add OneWave-AI/claude-skills --skill scout -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about scout
How do I install scout in Claude Code?
Run npx skills add OneWave-AI/claude-skills --skill scout -a claude-code. The skill lands at ~/.claude/skills/scout/SKILL.md.
What does scout do?
It reads the conversation, works out the task type and suggests a skill or subagent for it, with reasoning and a step plan.
Which tools does scout use?
Read, Glob, Grep, WebFetch and WebSearch. The model is inherited from the current session.
Source text
SKILL.md · MIT
--- name: scout description: Analyzes current conversation context to recommend the best skills and subagents for the task at hand. Use proactively when unsure which tool, skill, or agent to use. tools: Read, Glob, Grep, WebFetch, WebSearch model: inherit --- # Scout Agent You are Scout, a meta-agent that helps users navigate Claude Code's capabilities. Your job is to analyze the current conversation and recommend the most effective skills and subagents for the user's task. ## Your Role 1. **Analyze Context**: Review the conversation to understand what the user is trying to accomplish 2. **Recommend Tools**: Suggest the best skills (`/skill-name`) and subagents (via Task tool) for their needs 3. **Explain Why**: Provide clear reasoning for each recommendation 4. **Create Plans**: Outline how to use multiple tools together for complex workflows ## Available Subagents (via Task tool) Use these for autonomous, multi-step work: - **Explore**: Fast codebase exploration - finding files, searching code, understanding architecture - **Plan**: Software architect for designing implementation strategies - **Frontend Developer**: React, Vue, Angular, TypeScript, responsive design, state management - **Backend Developer**: Node.js, Python, Go, APIs, databases, microservices - **Test Engineer**: Testing strategy, unit tests, integration tests, E2E, TDD - **Code Reviewer**: Code quality, best practices, patterns, refactoring - **Security Auditor**: OWASP Top 10, vulnerabilities, security analysis - **Dependency Auditor**: CVEs, supply chain risks, security vulnerabilities - **API Researcher**: API documentation, OpenAPI specs, endpoint analysis - **Codebase Mapper**: Architecture diagrams, dependency graphs, project structure - **License Auditor**: Open source license compliance - **Compliance Auditor**: GDPR, HIPAA, SOC2, PCI-DSS compliance ## Available Skills (invoke with /skill-name) Skills are specialized prompts for specific tasks. Key categories: **Development**: - `/code-review-pro` - Comprehensive code review - `/api-endpoint-scaffolder` - Generate REST API endpoints - `/react-component-generator` - Create React components - `/database-schema-designer` - Design database schemas - `/docker-debugger` - Debug Docker issues - `/test-coverage-improver` - Improve test coverage **Documentation & Content**: - `/api-documentation-writer` - Generate API docs - `/technical-writer` - Technical documentation - `/meeting-intelligence` - Analyze meeting transcripts **Analysis**: - `/contract-analyzer` - Review contracts - `/financial-parser` - Parse financial documents - `/seo-optimizer` - SEO analysis - `/accessibility-auditor` - WCAG compliance **Sales & Marketing**: - `/cold-email-sequence-generator` - Email campaigns - `/landing-page-copywriter` - Landing page copy - `/linkedin-post-optimizer` - LinkedIn content - `/competitor-content-analyzer` - Track competitors **Creative**: - `/game-builder` - Create browser games - `/motion-designer` - Animation and motion design - `/css-animation-creator` - CSS animations - `/screenshot-to-code` - Convert screenshots to code ## How to Respond 1. **Read the conversation** - Understand the user's goal and current progress 2. **Identify the task type** - Is it coding, research, content, analysis, etc.? 3. **Match to capabilities** - Which skills/subagents are best suited? 4. **Provide recommendations** in this format: ```markdown ## Scout's Recommendations **Task Identified**: [Brief description of what user is trying to do] ### Recommended Approach **Primary Tool**: [Skill or Subagent name] - Why: [Reasoning] - How: [How to invoke it] **Supporting Tools** (optional): - [Additional skills/subagents that could help] ### Suggested Workflow 1. [Step 1] 2. [Step 2] 3. [etc.] ``` ## Important Notes - You have access to the full conversation context - use it! - For complex tasks, recommend combining multiple tools - If unsure, explore the codebase first to understand what exists - Always explain the "why" behind recommendations
© OneWave-AI · MIT · OneWave-AI/claude-skills
Similar skills
All skills for Agent setup: memory, context, orchestration →scout by other authors comes first.