agent_mcp_thoughtflow
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 | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| create_taskB | Create a new task with optional dependencies and metadata |
| get_taskB | Get a task by ID |
| list_tasksA | List all tasks, optionally filtered by status |
| update_taskC | Update a task |
| delete_taskC | Delete a task |
| create_workflowB | Create a workflow with a set of tasks |
| get_workflowB | Get a workflow by ID |
| list_workflowsB | List all workflows |
| delete_workflowC | Delete a workflow |
| create_strategyB | Create a strategy for organizing related workflows |
| get_strategyC | Get a strategy by ID |
| list_strategiesB | List all strategies |
| add_tree_to_strategyD | Add a tree to a strategy |
| remove_tree_from_strategyC | Remove a tree from a strategy |
| add_workflow_to_strategyB | Add a workflow to a strategy |
| remove_workflow_from_strategyC | Remove a workflow from a strategy |
| clear_allB | Clear all data (tasks, workflows, strategies, trees) |
| start_workflow_executionC | Start execution of a workflow |
| advance_workflow_runB | Advance a workflow run after task completion |
| get_workflow_runB | Get a workflow run by ID |
| list_workflow_runsC | List all workflow runs |
| delete_workflow_runC | Delete a workflow run |
| get_subtasksB | Get all subtasks of a parent task |
| move_taskB | Move a task to a new parent or change its order |
| deduplicate_strategiesA | Deduplicate strategies by normalized name. Keeps the first occurrence of each unique normalized name and removes duplicates. |
| create_treeC | Create a new Tree of Thoughts for systematic reasoning |
| get_treeB | Get a tree by ID |
| list_treesB | List all trees |
| delete_treeC | Delete a tree |
| add_ideaC | Add a child thought to an existing thought |
| get_thoughtC | Get a thought by ID |
| evaluate_thoughtC | Evaluate a thought with a score (0-100) and optional multi-criteria fields |
| verify_thoughtB | Mark a thought as verified after confirming its findings |
| select_thoughtC | Mark a thought as selected for further exploration |
| backtrackA | Backtrack from a thought, marking all descendants as pruned |
| prune_treeC | Prune thoughts below a certain evaluation threshold |
| clear_all_treesD | Clear all trees |
| deduplicate_treesA | Deduplicate trees by normalized goal. Keeps the first occurrence of each unique normalized goal and removes duplicates. |
| generate_children_with_llmA | Generate child thoughts using the configured LLM provider (Grok, Ollama, etc.). This triggers actual API calls to the LLM service. |
| promote_thought_to_tasksC | Promote a thought (or subtree) to executable tasks |
| spawn_tot_from_taskC | Spawn a Tree of Thoughts from a task for reasoning |
| link_thought_to_taskA | Create a soft bidirectional link between a thought and a task for provenance tracking |
| get_cognitive_provenanceB | Get cognitive provenance chain for a task or thought |
| deduplicate_strategies_and_treesA | Deduplicate strategies and trees by their normalized name/goal. This cleans up duplicate entries that may have been created before the deduplication logic was added. Keeps the first occurrence of each unique normalized name/goal and removes duplicates. |
| reload_stateA | Reload state from storage and share it across all services. Useful after manually deleting or editing the state file. |
| clear_stateA | Clear all state from storage and memory. This deletes the state file and resets all services to empty state. Useful for starting fresh without restarting the server. |
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 46 tools
There are several tools with overlapping purposes, such as the three deduplicate tools (deduplicate_strategies, deduplicate_trees, deduplicate_strategies_and_trees) and the three clear tools (clear_all, clear_all_trees, clear_state). While the descriptions help differentiate them, an agent might still struggle to select the correct one, especially when clear_all and clear_state appear to have similar effects. Overall, most tools are distinct, but these overlapping groups reduce clarity.
All tool names follow a consistent verb_first snake_case pattern. For example, add_idea, create_task, list_trees, delete_workflow_run. Even longer names like add_tree_to_strategy and generate_children_with_llm adhere to this pattern. There are no mixed conventions or confusing abbreviations.
With 46 tools, this server has a very large tool surface. The server aims to manage multiple domains (thoughts, tasks, workflows, strategies, trees), but the count is well above the recommended range (typically 3-15). Many tools could potentially be consolidated or removed without loss of functionality.
The tool set covers core CRUD operations for most entities but has notable gaps. For example, there are no tools to update or delete strategies, update or delete workflows (only delete_workflow exists but no update_workflow), or update thought content. Additionally, there is no tool to update a tree. These gaps may force agents to work around missing functionality.