Run Your First Skill

Run a Consuelo OS skill through the OS portal.

Run your first skill

A skill is an approved workflow an agent can run through Consuelo OS.

Use this page to verify the basic OS loop. The agent can reach OS. OS can describe available skills. A selected skill can run. The result is recorded with a trace ID.

Before you start

You need local OS installed, at least one connected agent, a passing health check, and one safe skill available.

Expected flow

agent asks OS for context
  -> OS returns skills and capability health
  -> agent selects a skill
  -> OS validates input
  -> OS creates an execution record
  -> script runs behind the skill
  -> OS returns a structured result

Expected result

A useful result includes status, skill name, trace ID, artifacts created, summary, and a safe error message when something fails.

If a capability is missing, OS should explain which capability is missing and how to add it.

Why this matters

This is the core loop. Every deeper workflow depends on the same path: context, skill, script, result, trace.

Related

• Skills
• Scripts
• Approvals
• Observability

Troubleshooting

Problem	Check
Agent cannot see OS	Reconnect the agent and restart the agent app.
OS returns no skills	Check skill registration and capability health.
Skill cannot run	Check the required capability list for that skill.
Result has no trace ID	Treat the execution as incomplete and inspect local logs.
Artifact missing	Check the artifact layer instead of looking only for a local file.

What good looks like

A first skill run should feel boring. The agent asks OS for context, OS shows a small set of approved skills, the selected skill runs, and the result includes enough metadata to debug it later.

The first skill does not need to prove every integration. It needs to prove the OS loop: context, skill, script, result, trace.