Loading ToolBench
MCP server for the Mukhtabir API, exposing interview, candidate, feedback, and webhook operations
30 tools with consistent naming (verb_noun pattern) and basic schemas. All tools have descriptions (10-50 chars, below ideal 50-200 baseline). Input schemas present but minimal, most parameters lack detailed constraints, ranges, or format guidance. No output schemas documented. Error handling absent. Missing parameter descriptions for optional fields (page, page_size). Tool composition is sound (single responsibility), but descriptions are too terse to guide LLM selection effectively. No evidence of idempotency markers, confirmation patterns for destructive ops, or recovery guidance.
Add evaluation criteria to a Mukhtabir interview.
Add a question to a Mukhtabir interview.
Add a subquestion to a Mukhtabir interview question.
Create a new Mukhtabir interview.
Create a Mukhtabir webhook.
Delete evaluation criteria from a Mukhtabir interview.
Descriptions too terse (10-50 chars vs. 50-200 baseline). Examples: 'Register a Mukhtabir candidate' (34 chars), 'List Mukhtabir candidates' (25 chars). LLMs cannot distinguish when to use similar tools (e.g., list_interviews vs. list_interview_results) without richer context.
Optional pagination parameters (page, page_size) lack descriptions. LLMs cannot infer valid ranges, defaults, or whether they are required. No min/max constraints visible.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., interview_id after create_interview). Responses are opaque.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | <=2025-11-25 | v2 |
Delete a question from a Mukhtabir interview.
Delete a subquestion from a Mukhtabir interview.
Delete a Mukhtabir webhook.
Fetch a Mukhtabir candidate by email address.
Fetch a Mukhtabir feedback record.
Fetch the recording URL for a Mukhtabir feedback record.
Fetch a Mukhtabir feedback transcript.
Fetch a Mukhtabir interview by ID.
Fetch analytics for a Mukhtabir interview.
Fetch a Mukhtabir webhook by ID.
Invite a candidate to a Mukhtabir interview.
List Mukhtabir candidates.
List paginated results for a Mukhtabir interview.
List Mukhtabir interviews.
List paginated deliveries for a Mukhtabir webhook.
List Mukhtabir webhooks.
Publish a Mukhtabir interview.
Register a Mukhtabir candidate.
Trigger a Mukhtabir webhook test delivery.
Update a Mukhtabir interview.
Update evaluation criteria in a Mukhtabir interview.
Update a question in a Mukhtabir interview.
Update a subquestion in a Mukhtabir interview.
Update a Mukhtabir webhook.
Destructive tools (delete_interview_question, delete_interview_criteria, delete_webhook) lack confirmation patterns or dry-run support. No error recovery guidance. Agents can accidentally delete without safeguards.
update_interview and update_webhook lack parameter descriptions for what fields can be updated. LLMs cannot determine valid inputs without trial-and-error.