skills.rest
Agent-Field
Agent-FieldOFFICIAL
Explore all skills by this author →
Website𝕏@agentfield_aiGitHub

agentfield-personal

Builds, packages, installs, and registers a personal AI agent on the local AgentField control plane.

2.5k|411|Updated Nov 5, 2025
Search Tags:#AI Agent#control plane#local deployment#cli install#agent packaging#Secrets Management#agentfield
One-click install
npx skills add https://github.com/Agent-Field/agentfield --skill agentfield-personal
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agentfield-personal
Source: https://github.com/Agent-Field/agentfield/tree/main/skills/agentfield-personal
Command: npx skills add https://github.com/Agent-Field/agentfield --skill agentfield-personal

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Turning an agent idea into a persistent, callable capability on your own machine normally requires manual packaging, installation, secret configuration, and registration steps that are easy to get wrong. This Skill walks through the full lifecycle so the agent ends up running, registered on the local control plane, and visible in AgentField Desktop.

Core Features & Use Cases

  • Source and packaging workflow: Authors real agent source in ~/agentfield-agents/<name> and packages it with an agentfield-package.yaml manifest declaring entrypoint, healthcheck, node ID, port, and dependencies.
  • Secret and lifecycle management: Declares user_environment secrets with global or node scope, configures them via af secrets set, then installs, starts, and verifies registration with af install, af run, and the control plane nodes API.
  • Live verification and handoff: Invokes the public entry reasoner through the control plane, diagnoses failures with af logs, and hands off restart, stop, and update commands.
  • Use Case: You want a pricing agent that lives on your machine and answers catalog questions. This Skill builds the source, declares the OpenRouter API key as a secret, installs and starts the agent, and confirms it answers a live request.

Quick Start

Build me a personal pricing agent on my local AgentField that answers product pricing questions and shows up in AgentField Desktop.

Frequently Asked Questions about agentfield-personal

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

FAQPage Schema
How do I install a personal agent on AgentField?▼

Author the agent source in ~/agentfield-agents/<name>, write an agentfield-package.yaml manifest, then run `af install` followed by `af run <name>`. Verify success by polling the control plane nodes API until the node registers as healthy.

What goes in an agentfield-package.yaml manifest?▼

The manifest declares config_version, name, release version, description, author, language, a runnable entrypoint.start command, a healthcheck path, the agent_node node_id and default_port, dependencies, and any user_environment secrets with their scope.

How do I configure API keys for an AgentField agent?▼

Declare each key in the manifest's user_environment with type secret and a global or node scope, then set values with `af secrets set KEY` for global keys or `af secrets set --node <name> KEY` for node keys. Secret values are never written into the manifest.

When should I build a personal agent instead of a standalone repository?▼

Choose a personal agent when you want a persistent capability on your own machine managed by the local control plane and visible in AgentField Desktop. A standalone repository with Docker Compose fits deployable projects and is covered by the agentfield skill.

Why is my AgentField agent installed but not working?▼

An install entry or spawned process alone does not mean success. Check `af logs <name>` first, then confirm the node appears as active in GET /api/v1/nodes and that a live reasoner invocation returns a successful result.

Related Skills

Explore More Software Engineering Skills→
Agent-FieldAgent-Field
2.5k

agentfield

Design and run multi-agent systems with AgentField's control plane.

Official
Advanced
az9713az9713
2

paperclip

Manage tasks and coordinate agents via the Paperclip control plane API.

Community
Advanced
octave-commonsoctave-commons
1

docker-compose

Orchestrate multi-container deployments defined in docker-compose.yml files.

Official
Intermediate
darelabs-techdarelabs-tech

dare-patterns

Mine brownfield software patterns via AST analysis to generate architectural documentation and JSON facts.

Official
Intermediate