Introduction & Quick Start
Relevant source files
The following files were used as context for generating this wiki page:
Scoutline is an agent-first Command Line Interface (CLI) for multimodal source investigation. It provides integrated tools for web research, repository exploration, and scholarly lookups with built-in provider routing, caching, and provenance. The system manages complex research tasks such as multi-provider web searches, PDF reading, site crawls, and temporal lookups via the Wayback Machine.
Sources: packages/scoutline/package.json, packages/scoutline/tests/agent-registration-docs.test.js:32-42
Core Capabilities
Scoutline organizes research tasks into specialized namespaces. Each namespace offers specific commands to interact with different types of data sources.
| Capability | Namespace | Primary Function |
|---|---|---|
| Web Research | search, read | Conducts multi-provider searches and fetches page content or PDFs. |
| Science | science | Performs scholarly lookups and searches across academic suppliers. |
| Repository | repo | Explores code repositories through search, file reading, and tree traversal. |
| Management | init, doctor | Handles agent registration and provider health checks. |
Sources: packages/scoutline/tests/science-docs.test.js:80-87, packages/scoutline/tests/agent-registration-docs.test.js:32-42, packages/scoutline/tests/repository-brief.test.js:15-20
Quick Start for Users
Installation and Setup
Scoutline requires Node.js version 24.0.0 or higher. You must configure API keys for specific providers in your environment variables to enable their capabilities.
- Install dependencies and build the project:
cd packages/scoutline
npm install
npm run build
- Initialize the CLI and register with agent tools:
scoutline init
- Verify your configuration and provider health:
scoutline doctor --health
Sources: packages/scoutline/package.json:28-30, 48-52, CONTRIBUTING.md:5-10, packages/scoutline/tests/agent-registration-docs.test.js:125-135
Common Research Workflows
The following diagram illustrates the typical data flow when performing a research task through the Scoutline CLI.
The diagram shows how the CLI dispatches requests to specialized providers, manages results through a central cache, and formats them for the user.
Sources: packages/scoutline/tests/agent-registration-docs.test.js:32-42, packages/scoutline/tests/repository-brief.test.js:630-645
Command Examples
Scholarly Search
Perform keyless scholarly lookups across academic suppliers like arXiv, OpenAlex, and PubMed.
scoutline science search "quantum computing"
scoutline science get <identifier>
Sources: packages/scoutline/tests/science-docs.test.js:70-87
Repository Exploration
Generate a brief overview of a repository including its structure, README, and manifests.
scoutline repo brief owner/repo --focus structure,readme
Sources: packages/scoutline/tests/repository-brief.test.js:630-640
Agent Registration
Register Scoutline with AI agents (e.g., Claude, Codex, Copilot) to provide them with research guidance.
scoutline init
# To reverse registration:
scoutline init --unregister
Sources: packages/scoutline/tests/agent-registration-docs.test.js:125-140
Development and Contribution
Developers can extend Scoutline by adding new "skills" or modifying provider adapters.
Project Structure
packages/scoutline: Contains the main CLI logic and source code.packages/scoutline/skills/: Contains marketplace skills shipped with the npm package.dist/: Target directory for compiled JavaScript code.
Sources: CONTRIBUTING.md:12-16, packages/scoutline/package.json:18-25
Adding a Skill
To add a new skill to the marketplace:
- Create a lowercase folder in
packages/scoutline/skills/. - Add a
SKILL.mdfile containing YAML front matter (name and description). - Include concise, action-oriented instructions for the agent.
Sources: CONTRIBUTING.md:18-24
Summary
Scoutline provides a robust framework for multimodal investigation by consolidating web, academic, and code-based research into a single CLI. It emphasizes reproducible results through mandatory caching and provenance, making it a primary tool for investigative research pipelines.