kata
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| KATA_PACKS_DIR | No | Overrides the packs directory. Defaults to ~/.claude/kata/packs. | |
| KATA_GLOBAL_DIR | No | Overrides the global chains directory. Defaults to ~/.claude/kata. | |
| KATA_PROJECT_ROOT | No | Overrides the project root directory. Defaults to the server process cwd. |
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": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_chainsA | List all available thinking chains: built-ins (master router, default freeform sequential thinking) plus file-defined chains from the global (~/.claude/kata), pack (/packs/, declared in /.claude/kata.json) and project (/.claude/kata) layers. Shadowing: project > pack > global. Also reports invalid chain files (with reasons), missing packs, pack conflicts and shadowing. |
| run_chainA | Start a thinking chain by name. When the task shape clearly matches a chain, call that chain directly; when nothing matches or the task needs multi-step reasoning, call run_chain("master") to route between PASS (no chains), "default" (freeform step-by-step thinking) and custom chains. Checklist chains return one complete checklist (single-shot). Staged/freeform chains open a session; continue with advance_chain. |
| advance_chainA | Submit the current stage's result and receive the next stage. Staged chains: pass exactly one of stage_output (work done for this stage) or skip_reason (why the stage was skipped). Freeform default chain: pass each thought as stage_output; set done: true with your final verified thought. expected_stage_index guards against retries: re-sending the previous index returns the same response idempotently. |
| save_chainA | Create or update a custom thinking chain as a Markdown file with YAML frontmatter (fields: name, description, mode: checklist|staged, optional domain/language/schema_version). The content is fully validated before writing; nothing is written on validation failure. name must equal the frontmatter name. Existing chains are only replaced when overwrite: true. |
| export_chainA | Export a chain's raw Markdown source for sharing (returns content, sha256, suggested filename). To import a shared chain, read its file and call save_chain with the content. If the name exists in both global and project scope, pass scope to disambiguate. |
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 5 tools
Each tool addresses a distinct stage of the chain lifecycle: listing, starting, advancing, saving, and exporting. There is no meaningful functional overlap between them.
All five tools follow the consistent verb_noun snake_case pattern: list_chains, run_chain, advance_chain, save_chain, export_chain. The naming is predictable and uniform.
Five tools tightly cover the core chain lifecycle without redundancy. This is a well-scoped size for a focused server.
The set covers listing, running, advancing, saving, and exporting chains, giving solid execution and authoring coverage. The main gap is the lack of an explicit delete/remove_chain operation, though save_chain can overwrite and this is a minor shortfall.