Codex Gemini Delegator V2
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_agent_backendsA | List available agent backends and check their health. |
| list_agent_runsA | List all agent runs, their states, and recent activity. Use this before creating new workers. |
| delegate_to_agentA | Delegate a new task to a Gemini worker. Creates a new Git worktree for isolated execution. |
| continue_agent_runA | Send follow-up instructions to an existing Gemini worker to continue its workstream. |
| get_agent_run_reportA | Get the full results, recent events, and the Git diff of an agent run. |
| apply_agent_runB | Merge the changes from the agent's worktree back into the main branch. |
| cleanup_agent_runA | Cancel a running agent and delete its temporary worktree. |
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 7 tools
Each tool targets a distinct action: listing backends, listing runs, fetching a report, delegating a task, continuing a run, applying changes, and cleaning up. There is no overlap or ambiguity between them.
All tool names follow a consistent verb_noun pattern in lower snake_case, e.g., list_agent_backends, delegate_to_agent, continue_agent_run. The variation between 'list' and 'get' is minor and conventional.
With 7 tools, the set is well-scoped for managing agent worker lifecycles. Each tool fills a specific niche without redundancy or unnecessary bloat.
The tools cover the full lifecycle of agent runs: creating (delegate), continuing, reading reports, listing, applying, and cleaning up. No critical operations are missing for the stated purpose.