hosted-agents

Provide sandboxed remote execution environments for coding agents.

Updated Apr 15, 2025
One-click install
npx skills add https://github.com/khrore/nix-config --skill hosted-agents-khrore
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hosted-agents
Source: https://github.com/khrore/nix-config/tree/main/dotfiles/common/.config/opencode/skills/hosted-agents
Command: npx skills add https://github.com/khrore/nix-config --skill hosted-agents-khrore

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Hosted agents solve the limits of local agent execution by moving coding workloads into remote, sandboxed environments that provide consistent setups, unlimited concurrency, and fast session startup so users get immediate, reproducible workspaces for coding, testing, and PR creation.

Core Features & Use Cases

  • Prebuilt Image Registry: Regularly built environment images that include cloned code, installed dependencies, and warmed caches to minimize time-to-first-token.
  • Snapshot & Restore: Filesystem snapshots taken at key points to enable instant follow-up sessions and quick rollback.
  • Warm Pool Strategy: Maintain pre-warmed sandboxes to serve sessions with near-zero startup latency.
  • Per-Session Isolation & Authorship: Isolated session state (per-session databases) and git commit attribution using the prompting user's identity for PRs.
  • Real-Time API Layer: WebSocket/streaming APIs to broadcast token streams, tool execution updates, and file change notifications to multiple clients.
  • Use Cases: Background coding agents that run independently of user devices, multiplayer debugging or PR review sessions, Slack/chat-triggered repo edits, and multi-client workflows (web, VS Code, Chrome extension).

Quick Start

Start a warm sandbox for repository myorg/frontend, restore the latest prebuilt image snapshot, and run the user's prompt with commits attributed to Alice Developer.

Frequently Asked Questions about hosted-agents

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

FAQPage Schema
How do I scale sandboxed coding agents to eliminate local resource contention?

Sandboxed coding agents scale by moving workloads into remote execution environments, providing consistent setups and unlimited concurrency. This eliminates local resource contention by isolating per-session state and maintaining warm pools for near-zero startup latency.

How does a warm pool strategy minimize session startup latency for coding agents?

A warm pool strategy minimizes session startup latency by maintaining pre-warmed remote sandboxes. These sandboxes serve incoming coding agent sessions instantly, bypassing the need to provision new environments and ensuring immediate access to reproducible workspaces.

Can I attribute git commits to the prompting user for PRs generated by background agents?

Yes, git commits can be attributed to the prompting user for PRs generated by background agents. The sandboxed execution environment isolates per-session state and applies the prompting user's identity to commits, ensuring proper authorship for automated PR creation.

What is the best way to broadcast token streams and file changes to multiple clients?

The best way to broadcast token streams and file changes to multiple clients is using a real-time API layer. WebSocket and streaming APIs broadcast token streams, tool execution updates, and file change notifications to multiple clients simultaneously for multiplayer workflows.

How do I instantly restore a coding agent's filesystem state for a follow-up session?

Filesystem state is instantly restored for follow-up sessions using snapshot and restore workflows. The system takes filesystem snapshots at key points, enabling instant session continuation and quick rollback without re-cloning code or reinstalling dependencies.

Does a prebuilt image registry reduce time-to-first-token for remote coding agents?

Yes, a prebuilt image registry reduces time-to-first-token for remote coding agents. The registry contains regularly built environment images with cloned code, installed dependencies, and warmed caches, minimizing setup time when spinning up new sandboxed sessions.