VSCode extension for sending code snippets and paths to Claude. Designed to work with wcgw MCP server.
This is a VSCode extension, not an MCP server. The source provided shows only package.json with VSCode command definitions, not MCP tool registration. No input schemas are visible, no parameter descriptions are present, and no structured output documentation exists. Tools are defined as VSCode commands with only titles in the contributes.commands array. This does not conform to MCP standards and appears to be a companion extension for a separate MCP server (referenced as 'wcgw mcp server' in the description). Without access to the actual MCP server implementation, and given only VSCode extension metadata, the tool definitions cannot be properly evaluated against MCP quality standards.
Check and Attach to Screen Sessions
Copy Selection with Full Context
Copy Terminal Selection with Full Context
Send Editor Selection to Application
Send Terminal Selection to Application
Toggle Screen Session Polling
No input schemas visible. All six tools lack JSON Schema definitions with parameter types, constraints, or validation rules. Per HARD SCORING RULES, schema score must be 0 for tools without visible input schemas.
Descriptions are minimal (under 30 characters). Tool descriptions are single short titles like 'Send Editor Selection to Application' with no context on WHEN to use the tool, WHAT it returns, or dependency hints.
No parameter documentation. VSCode command definitions do not declare input parameters, their types, constraints, or descriptions. Tools cannot be used by agents without knowing what inputs they accept.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 21 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 13 | - | v1 |
No output schema documentation. Return types, field names, and response structure are not specified. Agents cannot infer downstream tool chaining or data extraction needs.
Tool naming does not follow verb_noun convention. Names like 'toggleScreenPolling' and 'checkScreenSessions' use camelCase and are VSCode-idiomatic rather than action-verb-noun (e.g. 'toggle_polling', 'check_sessions'). This complicates LLM parsing.
No error handling guidance. No documentation of failure modes, retryable vs. fatal errors, or recovery steps (e.g., 'If no sessions found, ensure screen is running').
No indication of idempotency or write safety. Tool descriptions do not state whether calls are safe to retry, whether they modify state, or whether they are read-only (despite risk labels in the spec). 'toggleScreenPolling' is marked WRITE but has no description clarifying the effect.
This source is a VSCode extension manifest, not an MCP server definition. Tool definitions exist only as VSCode commands in package.json. The actual MCP server (referenced in the description as 'wcgw mcp server') is not provided for evaluation. Cannot fully assess MCP compliance without the server implementation.