Elenchus MCP Server

How to use

Describe what you want to verify to your AI assistant. The system will engage in a Verifier↔Critic loop to surface issues, challenge findings, and converge on a final verdict. You can request checks on specific files or modules and specify focus areas such as security, correctness, and edge-case coverage. Use prompts that clearly identify targets (for example, a path like src/auth or a module like payment processing) and the desired verification focus.

How to install

Prerequisites: ensure you have Node.js installed (v18 or newer) and a modern MCP client that can load stdio MCP servers. You should have an environment capable of running npm or npx commands from your shell.

# Optional: verify Node.js version
node --version
```
```bash
# Install and start the Elenchus MCP server via MCP client integration (example shown in client configuration)
# You will typically add the server config to your MCP client settings as shown in the quick start snippet

Configuring your MCP client

Add the following MCP server configuration to your client’s settings so your assistant can access Elenchus through the standard MCP channel.

{
  "mcpServers": {
    "elenchus": {
      "command": "npx",
      "args": ["-y", "@jhlee0409/elenchus-mcp"]
    }
  }
}

elenchus_start_session

Initialize a new verification session by specifying the target, requirements, working directory, and optional mode and differential settings.

elenchus_get_context

Retrieve the current session context, including files collected, issues, and recommended actions.

elenchus_submit_round

Submit a Verifier or Critic round with full analysis output and optional issue updates.

elenchus_end_session

End the session and produce a final verdict with a summary.

elenchus_get_issues

Query issues with optional status filtering to review current findings.

elenchus_evaluate_convergence

LLM-based evaluation of verification convergence quality.

elenchus_evaluate_severity

LLM-based severity classification for identified issues.

elenchus_evaluate_edge_cases

LLM-based validation of edge cases to ensure coverage.

elenchus_submit_llm_evaluation

Submit evaluation results from an LLM-based assessor.

elenchus_checkpoint

Create a checkpoint for session state to enable rollback.

elenchus_rollback

Rollback to a previous checkpoint.

elenchus_apply_fix

Apply fixes and re-run verification as needed.

elenchus_ripple_effect

Analyze ripple effects across dependencies when changes occur.

elenchus_mediator_summary

Generate a summary of mediator actions and interventions.

elenchus_get_role_prompt

Return the current role prompt for a given role (verifier/critic).

elenchus_role_summary

Summarize role configurations and compliance requirements.

elenchus_update_role_config

Update role-related configuration such as minimum compliance and strict mode.

elenchus_start_reverification

Initiate a re-verification pass on a target after fixes.

elenchus_save_baseline

Save a differential baseline for comparison in future analyses.

elenchus_get_diff_summary

Get a summary of diffs between baselines or revisions.

elenchus_get_project_history

Fetch historical verification data for auditing.

elenchus_get_cache_stats

Retrieve statistics on response caching.

elenchus_clear_cache

Clear cached verification results.

elenchus_get_pipeline_status

Check the status of the tiered verification pipeline.

elenchus_escalate_tier

Escalate to a higher verification tier as needed.

elenchus_complete_tier

Complete the current verification tier.

elenchus_get_safeguards_status

Query the status of quality safeguards.

elenchus_update_confidence

Adjust confidence levels for verification results.

elenchus_record_sampling_result

Record results from sampling in safeguards checks.

elenchus_check_convergence_allowed

Check if convergence criteria can be satisfied.

elenchus_set_compression_mode

Configure data compression mode for responses.

elenchus_get_optimization_stats

Obtain statistics on optimization of verification tasks.

elenchus_configure_optimization

Configure optimization settings for performance.

elenchus_estimate_savings

Estimate potential time or resource savings from optimizations.

elenchus_generate_roles

Generate dynamic roles for Verifier and Critic.

elenchus_set_dynamic_roles

Set dynamic role configurations for adaptive workflows.