agent-browser

Browse, search, and filter agents by type, model, and category in a browser interface.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/turbobeest/dev-system --skill agent-browser
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/turbobeest/dev-system/tree/main/skills/agent-browser
Command: npx skills add https://github.com/turbobeest/dev-system --skill agent-browser

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Opens a lightweight web UI for browsing and selecting agents during pipeline execution.

Core Features & Use Cases

  • Agent index: Compile an index of agents for easy search and filtering.
  • Dashboard: Browse, filter, and compare agents by category, model, and tier.
  • Selection persistence: Save user selections to a JSON file for downstream use.

Quick Start

Open the agent browser via the UI and select desired agents for the current phase.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I browse and select agents during pipeline execution?

Agent-browser provides a lightweight web UI to browse, search, and filter agents by type, model, and category during pipeline phases. Open the interface via the dashboard and select or deselect agents for the current run; selections persist to JSON for downstream use.

Can I filter agents by model type in the browser?

Yes, the agent browser filters agents by model (opus, sonnet, haiku), agent type (Pipeline, Specialist, All), and category, enabling targeted selection across your agent inventory.

What setup is required to use the agent browser?

Compile an agent index using scripts/compile-agent-index.py, which populates .claude/dashboard/agents.json and agent-selection.json. The single-file UI (agent-browser.html) and a gate hook (hooks/agent-browser-gate.sh) enable phase-based access and URL parameter defaults.

How do agent selections persist after I choose them?

Selected agents are saved to agent-selection.json, a JSON file maintained in .claude/dashboard/ that downstream tasks can read to apply your choices consistently across pipeline runs.

When should I use agent browser in my pipeline?

Deploy agent-browser at pre-phase and post-phase gates to let users centrally discover and select agents instead of managing scattered manual configurations, reducing friction in agent assignment.