feat: add Bright Data Rapid Agent for web intelligence and real-time web data#2
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a new custom GitHub Copilot agent, the 'Bright Data Rapid Agent', designed for real-time web research and scraping using the Bright Data MCP server, and updates the README to reflect this addition. The review feedback highlights a critical security vulnerability where the agent is over-privileged with unnecessary 'execute' and 'edit' tools, posing an indirect prompt injection risk. Additionally, the feedback points out an invalid model name ('claude-sonnet-4.6') and recommends wrapping the rules section in '' tags to maintain consistency with other agents in the repository.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| name: brightdata-rapid-agent | ||
| description: Real-time web research via Bright Data MCP (free tier) — search, discover, and scrape | ||
| model: claude-sonnet-4.6 | ||
| tools: ['execute', 'read', 'edit', 'search', 'brightdata/*'] |
There was a problem hiding this comment.
Security Vulnerability: Over-privileged Agent (Indirect Prompt Injection Risk)
The agent is configured with execute and edit tools. Since this agent's primary purpose is to perform web research and scrape external websites, granting it permissions to execute arbitrary shell commands and edit local files poses a significant security risk.
If a scraped website contains a malicious prompt injection payload, it could trick the model into executing arbitrary commands or modifying the repository files.
Recommendation:
Remove execute and edit from the tools list to adhere to the principle of least privilege.
| tools: ['execute', 'read', 'edit', 'search', 'brightdata/*'] | |
| tools: ['read', 'search', 'brightdata/*'] |
| --- | ||
| name: brightdata-rapid-agent | ||
| description: Real-time web research via Bright Data MCP (free tier) — search, discover, and scrape | ||
| model: claude-sonnet-4.6 |
There was a problem hiding this comment.
Correctness: Invalid Model Name
The model claude-sonnet-4.6 does not exist. To maintain consistency with the other agents in this repository (which use claude-sonnet-4.5), please update the model field.
| model: claude-sonnet-4.6 | |
| model: claude-sonnet-4.5 |
| ## Rules | ||
|
|
||
| - **Bright Data only for the web:** Never answer from training data or built-in fetch for current web facts; use MCP tools and cite tool output. | ||
| - Ground every factual claim in tool output; never invent data or URLs. | ||
| - Prefer fewer, higher-quality tool calls (free tier has monthly limits). | ||
| - Respect robots, terms of service, and privacy; do not bypass paywalls or authentication. |
There was a problem hiding this comment.
Consistency: Missing <rules> tags
To maintain consistency with the other agents in this repository (such as research-agent.agent.md) and the agent template (templates/agent-template.md), the rules should be wrapped in <rules> and </rules> tags.
| ## Rules | |
| - **Bright Data only for the web:** Never answer from training data or built-in fetch for current web facts; use MCP tools and cite tool output. | |
| - Ground every factual claim in tool output; never invent data or URLs. | |
| - Prefer fewer, higher-quality tool calls (free tier has monthly limits). | |
| - Respect robots, terms of service, and privacy; do not bypass paywalls or authentication. | |
| ## Rules | |
| <rules> | |
| - **Bright Data only for the web:** Never answer from training data or built-in fetch for current web facts; use MCP tools and cite tool output. | |
| - Ground every factual claim in tool output; never invent data or URLs. | |
| - Prefer fewer, higher-quality tool calls (free tier has monthly limits). | |
| - Respect robots, terms of service, and privacy; do not bypass paywalls or authentication. | |
| </rules> |
1 participant
Agent Submission
Agent Name: Bright Data Rapid Agent
Category: [ ] Orchestration [ ] Full-Stack [ ] Coding [ ] Testing [ ] Design [x] Productivity [ ] DevOps [ ] Debugging
Type: [x] New Agent [ ] Agent Improvement [ ] Bug Fix [ ] Documentation
Description
This contribution adds a GitHub Copilot custom agent that integrates with the Bright Data MCP server for live web access.
Live web research via Bright Data MCP — search, discover, and scrape with citations
Uses
discover,search_engine,scrape_as_markdown, and batch variants for search and page retrieval.The agent fills these gaps:
Configures Copilot to use the Bright Data MCP server for web search and page retrieval using the
discover,search_engine,scrape_as_markdown, and batch tools.Instructs Copilot to perform live web operations through the configured Bright Data MCP server and cite retrieved sources using the page title and URL.
The agents can be selected from the Copilot agent picker on GitHub.com, when assigning issues, in Copilot Chat (
/task), or through Copilot CLI (/agentor--agent).Testing Checklist
Documentation Checklist
Quality Standards
Additional Notes
Official references:
By submitting this PR, I confirm that: