CHANGE: Add Bright Data Rapid research chat mode

[felixLandlord]

Pull Request

Type of Change

• New Custom Chat Mode
• New Prompt File
• Update Existing Chat Mode
• Update Existing Prompt File
• MCP Server Configuration
• Documentation Update
• Bug Fix
• Other (specify below)

Asset Details

For New/Updated Chat Modes

• Chat Mode Name: brightdata-rapid-agent.chatmode.md
• Target Domain/Technology: Web research, Bright Data MCP
• Tools Required: discover, search_engine, scrape_as_markdown, search_engine_batch, scrape_batch, codebase, search

For New/Updated Prompt Files

• Prompt File Name: N/A
• Prompt Mode: N/A
• Target Use Case: N/A

Description

Adds a custom chat mode for live web research through the Bright Data MCP server (Rapid mode), plus a sample MCP server entry so users can wire it up easily.

Problem: Copilot often answers current-web questions from stale training data or built-in fetch, which can be blocked, incomplete, or uncitable.

Improvement: This mode requires Bright Data MCP for all live web search and page scraping, grounds claims in tool output with title + URL citations, and keeps workspace file access via codebase / search.

Useful scenarios:

• Recent company announcements and changelogs
• Cross-vendor pricing / feature comparisons from live pages
• Topic research with cited sources from the current week

Also registers the brightdata HTTP MCP server in mcp/mcp.json using a secure ${input:brightdata_api_token} prompt (no secrets committed).

Testing

• Validated chat mode front matter: quoted description, Rapid tool list + workspace tools
• Confirmed naming/location: chatmodes/brightdata-rapid-agent.chatmode.md
• Confirmed README entries in root README.md and chatmodes/README.md (alphabetical, install badges in root)
• Validated mcp/mcp.json JSON and that the token uses ${input:brightdata_api_token} (password input)
• Confirmed instructions utilizes the configured tools

Recommended local verification (optional):

1. Copy the brightdata server + brightdata_api_token input from mcp/mcp.json into workspace/user MCP config
2. Start the MCP server and provide a Bright Data API token when prompted
3. Select Bright Data Rapid Research chat mode
4. Run a search + scrape prompt and confirm citations come from Bright Data tool output

Examples

Sample Usage

Find the three most recent announcements from [company] and summarize each with a source link.

Compare current pricing across these two vendor pages: [URL 1] vs [URL 2].

Research what's currently being said about [topic] this week and cite sources.

Expected Behavior

• Uses Bright Data MCP tools for search/discovery and page scraping
• Scrapes promising URLs before stating page-specific facts
• Returns a synthesis with inline citations (title + URL)
• On MCP errors: retries once, tries an alternate Bright Data path, reports tool/URL/error, and returns any partial results
• May use codebase / search only for files in the current workspace

Quality Checklist

• I have read the contributing guidelines
• My contribution follows the repository's file naming conventions
• I have included proper YAML front matter (for .chatmode.md and .prompt.md files)
• I have tested my contribution in VS Code with GitHub Copilot
• My chat mode includes appropriate tools for its functionality
• I have provided clear examples and use cases
• I have updated the README.md if adding a new asset
• My contribution is self-contained and well-documented

Compatibility and Standards

• My contribution follows GitHub Copilot best practices
• I have considered accessibility and usability
• My content uses professional, inclusive language
• I have avoided proprietary or organization-specific information
• My contribution works across different development environments

Community Value

• Addresses unreliable “current web” answers without live, citable sources
• Useful whenever developers need fresh announcements, pricing, docs, or topic research inside Copilot chat
• Saves time vs manual browser hopping by combining search + scrape + synthesis in one mode
• Sample MCP config lowers setup friction for Bright Data’s generous free tier

Related Issues

N/A

Additional Notes

• Known limitation: Requires the Bright Data MCP server configured and running in the user’s workspace/user MCP config; mcp/mcp.json is a sample only
• Known limitation: Free-tier monthly usage limits apply — prefer fewer, higher-quality tool calls
• Security: API token is collected via VS Code input prompt and is not stored in the repository
• Scope: Rapid tools only; browser automation and web_data_* tools are intentionally out of scope for this mode as they are Paid services

---

This change is