Loading ToolBench
Lightweight, zero-dependency stdio MCP (Model Context Protocol) server for TimmyTest. Allows Cursor, Claude Code, Claude Desktop, Antigravity, and Zed to invoke TimmyTest tools natively with zero token waste and zero terminal scraping.
Server has 5 tools with complete input schemas and descriptions. Naming follows verb_noun pattern (timmytest_check, timmytest_scan, timmytest_run, timmytest_prompt, timmytest_integrate). Descriptions are present and contextual (avg ~150 chars), explaining WHEN to use each tool. However, parameter descriptions are minimal (e.g., 'Absolute or relative path' lacks format/constraint details), and output schemas are not documented in the visible code. Error handling and recovery guidance are absent. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles (4 READ_ONLY, 1 WRITE). Parameter defaults are reasonable but lack validation rules in descriptions.
Use when the user asks to test a repository, debug failing tests, or find likely missing tests. Run the existing suite locally, summarize failures and test gaps, and return an AI handoff prompt. No AI API call is made by TimmyTest.
Use only when the user asks to configure TimmyTest for this repository's coding agents. Write agent instruction files and TimmyTest configuration into the project.
Use when a coding agent needs compact context to fix failing tests or write missing tests. Generate a diagnostic handoff prompt from the repository analysis.
Use for a focused test run after a code change or when debugging failing tests. Execute the project's test suite locally and summarize failures with rule-based diagnostic leads.
Use when the user asks which code lacks tests, or when running tests is inappropriate. Statically report likely untested classes, functions, and missing test files without executing project code.
Output schemas not documented. Tools return results but LLMs cannot predict field structure, forcing trial-and-error parsing and wasting tokens.
Parameter descriptions lack format/constraint details. 'project_path' says 'Absolute or relative path' but does not specify if it must exist, what happens if missing, or if symlinks are followed. 'filter_pattern' lacks regex syntax or example patterns.
No tool annotations despite clear risk profiles. timmytest_integrate is WRITE (modifies project files) and timmytest_check/run are READ_ONLY but lack destructiveHint/readOnlyHint annotations. Agents cannot distinguish safe from risky tools.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
No error handling or recovery guidance. If project_path does not exist, timeout_seconds is invalid, or test suite fails to parse, tools likely return raw exceptions. Descriptions do not explain what to do on failure.
timmytest_integrate 'force' parameter lacks clarity. Description says 'appends; never overwrites' but does not explain what 'appends' means (file concatenation? YAML merge?) or what happens if config already exists and force=false.