research-workflow-studio
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| DB | Yes | Absolute path to the SQLite database. The browser service and MCP process must use the same database. | |
| WEB_URL | Yes | URL of the Research Workflow Studio web service, for example http://127.0.0.1:8831. |
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 |
|---|---|
| rws_project_listB | List projects and confirmed foundations. |
| rws_record_getB | Read one exact record or historical version. |
| rws_get_contextB | Read concise project/task context, completed facts, pending gaps and suggested actions. Suggestions are not execution authorization. |
| rws_code_findB | Find existing implementations before building new code. Returns exact identities and declared contracts; supplied check evidence is not independent verification. |
| rws_result_queryC | Find exact outputs across runs in this project. Retains producer, case, configuration and observation coordinates; never silently merges repeats. |
| rws_project_createC | Create an independent project and its initial confirmed foundation; no demonstration records are inherited. |
| rws_research_saveC | Save a research question or optional hypothesis; does not decide scientific validity. |
| rws_experiment_saveB | Save an experiment and bind its confirmed project foundation; does not execute it. |
| rws_code_registerB | Register an immutable code identity and explicit parent references; never reads or runs the path. |
| rws_run_registerB | Record an actual run and status. Registration never launches a program; producer bindings are immutable. |
| rws_artifact_registerB | Register an immutable output with its actual producer and configuration. For a legacy output with unresolved provenance, explicitly set provenance=unresolved with source evidence and a description; never invent a producer run. Does not read arbitrary files. |
| rws_protocol_registerC | Register immutable evaluation metric definitions, preprocessing, aggregation and missing-data policy. |
| rws_dataset_registerC | Register an immutable dataset version and exact sample IDs. |
| rws_conversation_saveC | Register an existing conversation reference; does not create or send a chat. |
| rws_evaluation_importB | Validate and register existing typed evaluation results with fixed protocol/code/input versions; does not compute metrics. |
| rws_comparison_saveC | Save exact comparison inputs and fixed/varying conditions without moving their original producers. |
| rws_analysis_saveC | Save observations, interpretation and limitations; manual analysis may omit code. |
| rws_report_create_revisionC | Create a report version with immutable foundation and transitive evidence snapshots. |
| rws_spec_saveB | Define a method-level contract before implementation. Saves a new draft or ready revision; implementations pin an exact ready revision. Cases/configurations belong to runs, not method contracts. |
| rws_usage_saveB | Reference precise artifacts in a target experiment, including cross-project reuse. Original producer remains unchanged. |
| rws_usage_removeC | Revoke a material usage without deleting source artifacts or other uses. |
| rws_overview_proposeB | Save a proposed foundation revision for human review; this does not apply it. |
| rws_overview_applyC | Human review only. MCP cannot approve its own suggestion; use the application review screen. |
| rws_overview_dismissD | Human review only. MCP cannot approve its own suggestion; use the application review screen. |
| rws_project_stages_setC | Human-only current-stage decision. Browsing and completed runs never advance stages. |
| rws_amendment_proposeB | Propose an audited metadata/source/lineage/recovery correction for an exact record. Does not rewrite the source or approve the correction. |
| rws_amendment_applyB | Human-only approval of an exact pending amendment. Raw records and previous report snapshots remain unchanged. |
| rws_amendment_dismissA | Human-only dismissal of an exact pending amendment. Keeps the proposal history without applying its patch. |
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 28 tools
Most tools target distinct research-workflow objects and lifecycle actions, with parallel propose/apply/dismiss and save/register families that remain mostly separable through their descriptions. The main ambiguity is between read-oriented tools such as record_get, get_context, result_query, and code_find, and between save/register/import semantics, but these are usually clarified by the tool text.
All tools share the rws_ prefix and generally follow a resource_action pattern such as project_list, experiment_save, and artifact_register. A few deviate, notably rws_get_context and the longer rws_report_create_revision, but the overall convention is predictable.
The 28-tool surface is heavy, but it reflects a broad research lifecycle with projects, experiments, code, runs, artifacts, datasets, protocols, reports, and review workflows. Most tools appear to earn their place, yet the count is large enough to increase selection burden and is above the rubric's typical well-scoped range.
The surface covers a wide CRUD-like and audited lifecycle: project creation/listing, context and record reads, experiment/code/run/artifact/dataset/protocol registration, result querying, analysis and report creation, usage tracking, and amendment/overview review. Minor gaps exist around explicit listing or browsing of some entity types, but read tools and context retrieval provide workarounds.