Loading ToolBench
Indian mutual-fund NAV (net asset value) data. Keyless access to daily NAVs and full NAV history for ~16,000 Indian mutual fund schemes (AMFI data).
Three tools with complete input schemas and descriptions. Naming follows verb_noun pattern (search_, get_, latest_). Descriptions are adequate (100-200 chars) and explain prerequisites. However, output schemas are not documented, LLMs cannot predict response structure. Parameter descriptions lack constraints (e.g., limit ranges, format specs). Error handling returns generic error objects without recovery guidance. No tool annotations (readOnlyHint, idempotentHint). Missing parameter validation rules in descriptions (e.g., 'limit must be 1-30').
Get a mutual-fund scheme's NAV (net asset value) history from MFAPI — fund house, category, ISIN, latest NAV, and the most-recent NAV points (newest first). Use search_schemes to find a scheme_code first (e.g. 118550). Keyless.
Fast single-value lookup of a mutual-fund scheme's latest NAV (net asset value) from MFAPI. Returns the most recent NAV, its date, and basic scheme metadata. Use search_schemes to find a scheme_code first (e.g. 118550). Keyless.
Find Indian mutual-fund scheme codes by fund name. Fetches the full MFAPI scheme list (~16,000 schemes) and filters client-side by a case-insensitive substring match on the scheme name. This is the way to obtain a scheme_code for get_nav_history and latest_nav. Note: this call downloads the full list (~1-2MB). Keyless.
Output schemas not documented. LLMs cannot predict response structure (fields, types, nesting). Responses include optional fields (fund_house, category, isin) with no schema definition.
Parameter constraints not documented in descriptions. 'limit' parameter lacks min/max bounds in text (code enforces 1-30 or 1-120, but descriptions say 'default 15, max 30' without stating minimum). LLMs may pass invalid values.
Error responses lack recovery guidance. Errors return bare objects like {error: 'scheme not found'} without suggesting next steps (e.g., 'Try search_schemes() first').
No tool annotations. Tools are read-only and idempotent but lack readOnlyHint/idempotentHint annotations in schema. LLMs cannot infer safety properties.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
search_schemes downloads ~1-2MB on every call (full scheme list). No caching, pagination, or incremental filtering. Wasteful for repeated searches.