Proof-MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TLA2TOOLS_JAR | No | Path to the TLA+ tools JAR used for TLC checks. Omit when TLC checks are not needed. | |
| PROOF_MCP_WORKSPACE | No | The workspace for the MCP server. Set this if the server's working directory is not the project root. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
Proof-MCP supports agent-driven source-to-spec QA. For verification requests, discover scope and candidate commands, obtain user approval before running build/test commands, and prepare bounded source context for the selected module.
Draft candidate Lean 4 theorems for precise functional properties and TLA+ models for bounded stateful, concurrent, or temporal behavior. Call proof_check_generated to check generated source; users do not need checked-in .lean or .tla files. Keep generated artifacts temporary unless asked to retain them.
Report observations, assumptions, bounds, fairness, checker results, and uncovered behavior. A passing generated artifact does not prove source-to-model refinement or whole-program correctness.
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 |
|---|---|
| proof_check_leanA | Check one workspace-relative Lean 4 file with warnings treated as errors. |
| proof_run_python_testA | Run one workspace-relative Python test file. This executes local code and is runtime evidence, not a formal proof. |
| proof_check_tlaC | Run TLC on a workspace-relative TLA+ module and TLC configuration. |
| proof_discover_qaA | Inspect workspace manifests and source extensions to report detected languages, modules, and candidate build/test commands. This is planning only and never executes project commands. |
| proof_run_approved_checksA | Execute only discovered command IDs that the user explicitly approved. This runs local repository code. Never call this tool until the user has seen the plan and approved the selected IDs. |
| proof_prepare_verificationB | Discover source files in an optional scope, collect a bounded source context and Clarity dependency graph, and return instructions for the agent to draft properties and candidate Lean/TLA+ artifacts. |
| proof_check_generatedA | Check agent-drafted Lean and/or TLA+ content without requiring formal files in the repository. Optionally retain the checked bundle and its properties/assumptions under .proof-mcp/generated/. |
| proof_verify_workspaceA | Run artifacts from .proof-mcp.json. If no manifest exists, use proof_check_generated to check candidate artifacts without adding Lean or TLA+ files to the project. |
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 8 tools
Most tools have clearly distinct roles across discovery, preparation, execution, and verification. The main overlap is between proof_verify_workspace and proof_check_generated, and between proof_discover_qa and proof_prepare_verification, but descriptions clarify their boundaries.
All tools use a consistent proof_<verb>_<object> pattern, such as proof_check_lean, proof_run_python_test, and proof_verify_workspace. The shared prefix and verb-noun structure make the set predictable.
Eight tools is well-scoped for a verification workflow covering discovery, preparation, checking, and approved execution. Each tool has a distinct stage or artifact type, with no obvious padding.
The surface covers the main lifecycle from QA discovery through Lean/TLA+ checks and generated artifact verification. Minor gaps exist around inspecting prior results or managing .proof-mcp.json artifacts, but agents can work around them.