capsule-agents
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| CAPSULE_MODEL | Yes | The name of the Ollama model to use for real agent operations. Must be an installed model. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| capsule_runB | Run an isolated main/sub-agent team on explicit task and text copies. Requires Docker and a host-configured Ollama model. |
| capsule_demoB | Deterministic demo in actual Docker capsules; does not test a real model. |
| capsule_doctorB | Read-only Docker image readiness check. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
Each tool has a clearly distinct purpose: capsule_demo (deterministic demo), capsule_run (actual agent execution), and capsule_doctor (readiness check). No overlap in intent, so an agent can select correctly without ambiguity.
All three tools use a consistent capsule_ prefix followed by a short, lowercase suffix (demo, run, doctor). The pattern is predictable and uniform across the set.
Three tools is on the lean side but appropriate for a focused capsule-execution server, and each tool clearly earns its place. It is slightly thin rather than excessive.
The surface covers demo, execution, and readiness checking, but there is no way to inspect run status, retrieve results separately, or stop a running capsule. These are notable lifecycle gaps an agent may hit.