Skill as a Service

Spawn cloud code agents with specialized skills and execute tasks remotely. Specify the task, agents, and skills; the cloud handles provisioning and execution.

Agent Instructions: Setup & Authentication

Step 1: Verify API Key

Before performing any action, check if the SKILL_SERVICE_API_KEY environment variable is available.

• •If missing: Pause and ask the user: "I need a Rebyte API key to proceed. You can get one at rebyte.ai/settings/api-keys. Please provide your key or set the SKILL_SERVICE_API_KEY environment variable."
• •If present: Continue with the user's request.

Quick Start

1. Common Operations

You can use the Python client or the provided CLI tool.

Python Client

from scripts.skill_service import get_client

client = get_client()

# List available skills
skills = client.list_skills(search="pdf")

# Spawn an agent
agent = client.spawn_agent(
    agent_name="my-worker",
    skills=["anthropics-pdf"],
    prompt="You are a PDF expert."
)

# Run a task (waits for completion by default)
result = client.run_task(
    agent_id=agent["agent_id"],
    task_description="Extract text from the PDF",
    input_data={"file": "sample.pdf"}
)

# Terminate when done
client.terminate_agent(agent["agent_id"])

CLI Tool

# List skills
python3 scripts/skill_cli.py list-skills --search pdf

# Spawn an agent
python3 scripts/skill_cli.py spawn-agent --name worker --skills anthropics-pdf

# Run a task
python3 scripts/skill_cli.py run-task --agent-id AGENT_ID --task "Extract text" --input '{"file": "sample.pdf"}'

Detailed Resources

• •API Reference: See references/api.md for full method signatures, response formats, and error handling.
• •Advanced Examples: See references/examples.md for complex workflows, custom polling, and bulk processing.

Configuration