skills.rest
Apra-Labs
Apra-LabsOFFICIAL
Explore all skills by this author →
WebsiteGitHub

fleet

Register, authenticate, and dispatch commands to AI agent members across environments.

65|10|Updated Feb 25, 2026
Search Tags:#permissions#agent orchestration#remote execution#member onboarding#task monitoring#secure credentials#vcs authentication
One-click install
npx skills add https://github.com/Apra-Labs/apra-fleet --skill fleet-apra-labs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fleet
Source: https://github.com/Apra-Labs/apra-fleet/tree/main/skills/fleet
Command: npx skills add https://github.com/Apra-Labs/apra-fleet --skill fleet-apra-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of reliably running multi-agent work by provisioning and governing fleet “members” (local or remote), securely managing credentials and permissions, and dispatching commands or LLM prompts in a controlled way.

Core Features & Use Cases

  • Member lifecycle management: register, list, inspect details, update metadata, and remove members.
  • Secure onboarding & auth provisioning: set up SSH auth for remote members, provision LLM auth, and provision/revoke VCS credentials (GitHub/Bitbucket/Azure DevOps) with reusable secret storage.
  • Permissions and safe dispatch: compose provider-native permission configs, enforce denial handling, and manage unattended execution modes safely.
  • Work execution & orchestration: run shell commands (execute_command) and dispatch agent reasoning tasks (execute_prompt) with timeouts, resume behavior, concurrency guards, and cancellation via stop_prompt.
  • File transfer & persistence: push/pull files between orchestrator and members, and track long-running work using Beads (bd) for durable session/task recovery.

Quick Start

Ask the AI to register two members for a project, compose permissions, then execute a plan-to-implementation flow by dispatching the doer via execute_prompt and the reviewer via execute_prompt.

Frequently Asked Questions about fleet

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

FAQPage Schema
How do I coordinate remote execution across multiple AI agents?▼

Multi-agent orchestration works by registering local or remote members, composing provider-native permission configurations, and dispatching work via execute_command for shell commands or execute_prompt for LLM reasoning tasks with concurrency guards and timeouts.

How do I securely manage credentials for remote agent dispatch?▼

You securely manage credentials by provisioning SSH keys for remote members, configuring LLM provider authentication, and storing VCS credentials for GitHub, Bitbucket, or Azure DevOps using reusable secure secret handles.

Can I use this approach to run unattended automated code review loops?▼

Yes, you can automate code review loops by dispatching a doer agent via execute_prompt to implement changes and a reviewer agent to evaluate them, managing the entire process with unattended execution modes and enforced denial handling.

How do I handle timeouts and cancellation for long-running agent tasks?▼

You handle long-running tasks by configuring timeouts, resume behavior, and concurrency guards during execution, then issuing cancellation via stop_prompt to safely terminate work and tracking durable session recovery using Beads.

Does this method support provisioning VCS authentication for Bitbucket and Azure DevOps?▼

Yes, VCS authentication provisioning supports Bitbucket and Azure DevOps alongside GitHub, allowing you to provision and revoke credentials with reusable secret storage for secure multi-agent software sprints and infrastructure surveys.

Related Skills

Explore More Software Engineering Skills→
phnx-labsphnx-labs
15

devices

Synchronize SSH device profiles from Tailscale for fleet-wide command execution.

Official
Advanced
jayhjenkinsjayhjenkins

executing-plans

Execute implementation plans in batches with human review checkpoints.

Community
Intermediate
YH-05YH-05
1

agent-teams-file-passing

Validate file-based data transfer patterns for Agent Teams over 50KB.

Community
Advanced
himanshu231204himanshu231204
2

autonomous-agent-harness

Configure persistent self-directing agent workflows with Claude Code and MCP servers.

Community
Advanced