Loading ToolBench
MCP server for job search assistance, resume generation, and interview preparation. Provides tools to scrape LinkedIn jobs, generate resumes in PDF format, and generate interview questions.
Three tools with mixed quality. scrapeLinkedIn and generateResume have reasonable schemas but descriptions lack clarity on when/why to use them. interviewQuestions has a confusing schema where 'questions' is both an input parameter AND output field. All tools lack error handling guidance, recovery paths, and actionable error messages. Parameter descriptions are present but generic. No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Tool names lack clear verb-noun structure (scrapeLinkedIn uses camelCase; should be snake_case per MCP conventions). Output schemas are not documented.
Generates a resume from the given properties and returns some JSON data which is then converted to a downloadable PDF for the user
If the user ask for a list of resume questions then provide them with it so that they can practice before an interview. Make these questions relevant to the job descripion.
Scrapes LinkedIn for jobs based on the input that was given.
interviewQuestions has confusing schema: 'questions' is both an input parameter (array of strings) and appears to be the output. Schema is self-referential and unclear whether the tool generates questions or validates them.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields without knowing what each tool returns. generateResume claims to return 'JSON data which is then converted to a downloadable PDF' but structure is undefined.
Tool names use camelCase (scrapeLinkedIn, generateResume, interviewQuestions) instead of snake_case convention. MCP spec and production tools standardize on snake_case for consistency.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
No error handling or recovery guidance. If LinkedIn scraping fails, if resume generation fails, or if job description is invalid, tools provide no actionable error messages or next steps for the LLM.
scrapeLinkedIn description contains prescriptive instructions ('DO NOT USE THE TOOL IF YOU DON'T HAVE...') that should be enforced via schema constraints (required fields, validation) rather than natural language warnings. LLMs do not reliably follow imperative instructions in descriptions.