OmniForge MCP server exhibits pervasive definition quality gaps across 26 tools. While tool names follow verb_noun convention (omniforge_run_*, omniforge_skill_*, omniforge_workflow_*), descriptions are generic and lack LLM-optimized guidance. Critical issue: input schemas are severely underdeveloped. Most tools accept only 'parameters: {type: object, description: "Optional skill parameters"}' with NO nested property definitions, violating the constrained-input pattern. This forces LLMs to guess valid parameter names and values. Output schemas are completely undocumented, no tool declares what it returns. Error handling is absent; no recovery guidance or categorization. Security concerns: no evidence of permission gates, audit trails, or secret injection patterns. Composition is poor: 26 tools with overlapping responsibilities (e.g., omniforge_run_skill + 14 individual skill tools) create decision paralysis for LLMs.

• Expand skill tool input schemas from bare 'parameters: {type: object}' to explicit nested properties. E.g., omniforge_skill_test_driven_development should declare: testFramework (enum: jest|vitest|mocha), targetCoverage (number 0-100), includeIntegration (boolean). Reference constrained-input pattern.
• Document output schemas for all 26 tools. Specify: success response structure (e.g., {status: string, skillId: string, result: object, executionTime: number}), error response structure, and pagination if applicable. Enable tool-chain composition.
• Enhance descriptions to 100-200 chars with WHEN-to-use context. E.g., 'Execute a TDD workflow: write failing test, implement code, refactor. Use when starting a new feature or fixing a bug. Returns test results and coverage metrics.' Clarify distinction between omniforge_run_skill and individual skill tools.
• Add error handling with recovery guidance. Return structured errors: {code: string, message: string, retryable: boolean, suggestion: string}. E.g., 'Skill execution timeout. Retryable: true. Suggestion: increase timeout or break task into smaller steps.'
• Consolidate tool interface: remove redundant omniforge_run_skill/omniforge_run_workflow if individual skill/workflow tools exist, or vice versa. If both are needed, document in descriptions when to use each. Reduce cognitive load on LLM tool selection.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry