Prompt Writing Best Practices
To get the most out of HighQ MCP, we have curated some of the useful prompt categories, advanced use cases, and best practices.
Recommended | Not Recommended |
|---|---|
Show me documents in the Corporate/Active Matters folder uploaded in the last 30 days. Specific location, date filter, reasonable result size . | Show me a list of all sites in the HighQ instance I have access to. Need to search hundreds or thousands of sites, depending on the HighQ user-role, hits size limits . |
Get items from Matter Tracker where Status='Active' and Practice Area='M&A', limit to 50 records. Filtered query, explicit limit . | Show me all documents in the Corporate site. Hundreds or thousands of results, hits size limits . |
Analyse the top 10 most recently modified documents in the Litigation-2025 folder. Limited scope, specific criteria . | Get all items from the Matter Tracker iSheet. May return 500+ rows, truncates response . |
Count how many documents are in each subfolder of Real Estate, then show me the 3 largest folders. Aggregation first, then limited detail . | Analyse everything in my HighQ account. Too broad, requires too many tool calls . |
Search for 'lease agreement' in Real Estate site, return only documents from Q4 2025, show first 20 results. Specific search, date filter, result limit . | Download all files from the Litigation folder and summarise them. May exceed response size and token limits . |
Advanced use cases
- Create an interactive Acme co. Legal Portfolio Management System that: (1) displays all 10 franchise locations with their managers, addresses, and property values on a visual dashboard, (2) analyzes all guidance documents to extract legal compliance requirements and maps them to each location based on risk level, (3) generates a comprehensive General Counsel report showing total portfolio exposure, highest-risk properties, and regulatory obligations, and (4) includes a "What-If Analyzer" where I can simulate adding new acquisitions to see real-time impact on portfolio composition, risk profile, and legal resource allocation — make it a fully functional web application with interactive charts, filtering capabilities, and executive summary export.This type of advanced prompt demonstrates the full power of HighQ MCP combined with Claude's ability to generate interactive applications from real legal data. Claude will use multiple tools to fetch site data, documents, and iSheet records, then synthesise them into a functional dashboard.
- Create a table of documents with columns for duplicates, language, version numbers, signed yes or no, document types and check for any reading errors and flag any documents with a red mark that require M&A Legal review as they hold M&A sensitive provisions in project blue site.It instructs the system to search the Project Blue site, analyse its documents, and compile a table that identifies duplicates, language, version numbers, signature status, and document types, while also checking for readability/parsing errors. It also asks to flag (red mark) any documents that contain M&A sensitive provisions, so they're routed for M&A Legal review.
Best Practices
- Always define the scope (where to look).
- Add filters early (date range, metadata, keywords).
- Set an explicit result limit (top N / first N).
- Use aggregation first (counts/grouping) before listing everything.
- Break large requests into stages (find → shortlist → analyse).
- Specify the output format (table columns, bullets, CSV, etc.).
- Define key terms and rules (risk levels, relevance, priority order).
- Ask for evidence links when extracting requirements.
- Prefer exact phrases when needed ("…") and clear keywords.
- Keep prompt's objective and actionable (one task per prompt).