MCP server for Plan Forge — an AI-driven planning and execution framework with tool metadata, workflow definitions, and multi-model consensus capabilities
Plan Forge MCP exhibits severe definition quality gaps across all 12 tools. All tools lack input parameter descriptions, a critical violation of pattern:tool-description. Most tools have empty input schemas ({}), providing zero guidance to LLMs about what parameters are required or optional. Tool descriptions are present but generic and lack actionable context (e.g., 'Diagnose, inspect, and health-check the Forge setup' does not explain WHEN to call forge_smith vs forge_validate). Only 2 of 12 tools (forge_diff, forge_analyze, forge_run_plan) have documented input parameters with types; the remaining 9 have empty schemas. No tool documents output structure, making it impossible for LLMs to plan downstream calls. Error handling is absent, no recovery guidance, no categorization of retryable vs fatal errors. Security considerations (credentials, permissions, audit trails) are not visible in tool definitions.
Stop and cancel an active plan execution
Analyze plan consistency, traceability, coverage, and tests; return audit score
Generate cost and token usage report across runs and models
Delegate code review to GitHub Copilot by creating an issue on an open PR
Compare plan against current state and detect scope drift
Classify staged git changes by category (plan, test, docs, config, chore, scope, unknown)
Read and display the status and progress of plan executions
9 of 12 tools have empty input schemas ({}), violating pattern:constrained-input and pattern:tool-description. LLMs cannot determine what parameters are valid or required.
No tool documents output schema or return structure. LLMs cannot plan downstream tool calls or extract required fields for chaining (pattern:tool-chain violation).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
Execute a plan with multi-model consensus, slice-by-slice automation, and guardrails
Diagnose, inspect, and health-check the Forge setup
Read and display deployment roadmap status and phase overview
Scan and audit for completeness, find TODO markers
Validate and verify Forge setup configuration
Tool descriptions lack actionable context. 'Diagnose, inspect, and health-check' does not explain WHEN to call forge_smith vs forge_validate vs forge_sweep. Descriptions should state purpose, prerequisites, and when to use (pattern:tool-description).
No error handling guidance. Tools do not document what errors are retryable, what user action is needed, or how to recover (pattern:recovery-guide, pattern:error-classification violations).
forge_run_plan accepts 'quorum' enum with value 'false' (string), which is semantically confusing. Should be a boolean or use 'disabled'/'none'. Parameter descriptions also lack format/constraint details (pattern:constrained-input).