tree-sitter-analyzer
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TREE_SITTER_OUTPUT_PATH | No | Optional path for large output write target. | |
| TREE_SITTER_PROJECT_ROOT | Yes | Absolute path to the project root directory. |
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
} |
| prompts | {
"listChanged": true
} |
| resources | {
"subscribe": true,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| searchB | Code-intelligence (codegraph-compatible) search facade; multi-hop graph questions (callers then tests, untested callers, impact, depth/path filters) take ONE action=uql call with completeness. Actions: chain, query, select, semantic, subscribe, symbol, text, tql_execute, tql_schema, unsubscribe, uql. Pass action=help with target_action for an exact action schema. |
| navA | Code-intelligence (codegraph-compatible) navigation facade (multi-hop: search action=uql). Actions: call_path, callee_tree, callees, caller_tree, callers, co_change, context, impact, lineage, navigate, pulse, pulse_batch, resolve, test_map, trace, xref. Pass action=help with target_action for an exact action schema. Cost: action=callers is slow (prefer nav action=context (1-hop, warm)). |
| structureB | Code-intelligence (codegraph-compatible) structural analysis facade. Actions: analyze, ast_path, class_detail, class_tree, explore, outline, read, signatures, sitemap. Pass action=help with target_action for an exact action schema. |
| healthA | Code-intelligence (codegraph-compatible) health and analysis facade. Actions: dead, deps, file, heatmap, imports, matrix, middleware, overview, patterns, project, refactor_queue, routes, scale, self, test_gap, unreachable. Pass action=help with target_action for an exact action schema. Cost: action=project is slow (prefer health action=file (one file, warm)). |
| editB | Code-intelligence (codegraph-compatible) safety and change-management facade. Actions: ast_diff, classify, constraints, guard, impact, mutation_probe, plan_rename, pr, refactor, release_snapshot, rename, safe, verify. Pass action=help with target_action for an exact action schema. Cost: action=safe is slow. |
| projectC | Code-intelligence (codegraph-compatible) project-intelligence hub. Actions: card, doc_sync, journal, metrics, overview, parser, skills, smart, workflow. Pass action=help with target_action for an exact action schema. |
| indexB | Code-intelligence (codegraph-compatible) index lifecycle hub. Actions: auto, build, cache, full, knowledge, schema, status, sync. Pass action=help with target_action for an exact action schema. |
| vizB | Code-intelligence (codegraph-compatible) visualization and similarity facade. Actions: graph, knowledge, similarity, uml. Pass action=help with target_action for an exact action schema. |
| set_project_pathA | SMART Workflow 'Set' step (FIRST): Set the project root path for security boundaries. Call this before any other tool to ensure correct file resolution and security validation. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| smart_analyze | SMART Workflow: Systematic code analysis for a single file. Recommended for files you haven't seen before. |
| smart_explore | SMART Workflow: Explore a new project. Get the full picture before diving into code. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| code_file | Access to code file content through URI-based identification |
| project_stats | Access to project statistics and analysis data |
| Hyphae selector result | Re-evaluate a Hyphae selector expression and return the current result set. URI produced by search action=subscribe. |
TDQS
Scored across 9 tools
Eight facade tools bundle many actions, and their boundaries are porous: search and nav both advertise multi-hop/impact/UQL, nav and edit both expose impact, viz and index both expose knowledge, and health/project both expose overview. An agent can often justify several entry points for the same question, so misselection is likely despite detailed action lists.
Most tools are single-word noun facades (viz, search, nav, edit, health, index, project, structure), but set_project_path is a verb_noun setup tool and abbreviations like viz/nav break the pattern. The convention is mixed but still readable.
Nine top-level tools is a sensible count and the facade grouping keeps the surface navigable. However, each facade hides 8-16 actions, so the effective operation count is much larger than nine and some facades are overstuffed.
The surface covers index lifecycle, search/navigation, structural analysis, health, edit/refactor, visualization, and project intelligence, which is broad coverage for a code-intelligence server. No obvious lifecycle dead end exists, though some operations such as test execution or documentation generation are only implied or partial.