An AI agent platform with tool registry, external MCP integration, and agent templates for financial analysis, compliance, and research tasks
Helix is a Next.js application, not an MCP server. No MCP protocol implementation detected. Tools are registered in a custom tool-registry.ts with minimal schema documentation. Only 2 of 36 tools have visible input schemas (india_iip_index, web_search, todo_write). Most tools lack parameter descriptions, type definitions, and output schemas. Descriptions are present but generic (10-50 chars). No error handling guidance, no permission gates, no audit trails. This is a monolithic agent framework, not a production-grade MCP server.
Calculate acquirer capacity score
Assemble an investment banking pitch book
Get company information from Crunchbase
Search the data.gov.in catalog for datasets
Fetch data from data.gov.in datasets
Fetch and parse content from a URL with SSRF protection and 50KB size limit
Read messages from the agent's inbox
Get GDP series data from data.gov.in
Not an MCP server, this is a Next.js monolithic application with a custom tool registry, not a Model Context Protocol implementation. Cannot be used as an MCP server by hosted clients or remote agents.
34 of 36 tools have NO visible input schemas. Only india_iip_index, web_search, and todo_write show parameter definitions. Remaining tools cannot be validated or properly invoked by LLMs.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 23 | 2025-06-18+ | v2 |
Index of Industrial Production (IIP), monthly. Filter by sector (Mining / Manufacturing / Electricity / General). Source: MoSPI via data.gov.in. Use for industrial-sector benchmarking.
Extract data from insurance factsheets
Fetch insurance factsheets for Indian insurance companies
Get Consumer Price Index (CPI) data from MoSPI via data.gov.in
Get Wholesale Price Index (WPI) data from MoSPI via data.gov.in
Get RBI foreign exchange reference rates from data.gov.in
Get RBI policy rates from data.gov.in
Get SEBI mutual fund AUM data from data.gov.in
Delete from the agent's learning memory
Read from the agent's learning memory
Write to the agent's learning memory
Get historical financial data from Macrotrends
Allows an agent to modify its own system prompt
Extract details from precedent transactions
Search for precedent transactions
Execute code in a sandboxed environment
Screen companies based on criteria
Search the agent's knowledge base
Search for companies in SEC EDGAR database
Retrieve SEC EDGAR filings for a company
Extract XBRL facts from SEC EDGAR filings
Send a message from the agent
Spawn a sub-agent to perform a task
Record/update a research task list for complex multi-topic queries. Returns { completedCount, totalCount }.
Generate a valuation football field chart
Search the web. Returns ranked results, each with an [n] citation index, title, url, and snippet. Cite facts using these [n] indices.
Get financial data from Yahoo Finance
Get stock quote data from Yahoo Finance
Most tool descriptions are generic and under 50 characters (e.g., 'Read messages from the agent's inbox'). Lack context for LLM selection, prerequisites, or when to use vs. similar tools. Descriptions should be 50-200 chars with WHAT, WHEN, and WHY.
No error handling guidance. Tools like run_code (WRITE risk) and learning_memory_delete (DESTRUCTIVE) lack recovery hints, retry classification, or confirmation steps. Agents cannot self-correct on failure.
No permission gates or audit trails. Tools like modify_own_prompt and spawn_subagent can be invoked without scope checks or logging. No evidence of least-privilege configuration or compliance tracking.