Skip to main content
Alphabetized reference of Verdent-specific terms and technical concepts.

Core Concepts

The main AI assistant that processes requests and executes tools
When conversation history fills context window, limiting the ability to reference earlier content
Amount of conversation history the AI can reference
System for extending Verdent with external tools
Specialized AI agent with isolated context for specific tasks (Explorer, Verifier, Code-reviewer)
Function the agent can execute (file_read, bash, glob, etc.)

Execution Modes

Tools execute automatically without approval prompts
User approves each tool execution before running
Agent creates plan before executing changes
Run without permission prompts (advanced)

Files & Configuration

Project-specific team standards file (project root, version controlled)
MCP server configuration (~/.verdent/mcp.json)
Plan Mode output customization (~/.verdent/plan_rules.md)
Global personal preferences file (~/.verdent/user_rules.md)
Current VS Code project directory where Verdent operates

Tool-Specific Terms

File search pattern using wildcards (**/*.ts, src/**/*.js)
Specific lines to read from large files (start_line, max_lines)
Regular expression for advanced text searching
file_edit flag to replace multiple identical strings
Maximum time for bash commands: 120 seconds (2 minutes, hard limit)

Subagent Terms

Subagent can activate automatically based on request patterns
Rules for when subagent activates (strict vs flexible)
Subagent’s separate context window (doesn’t consume main context)
Subagent only runs when explicitly @-mentioned
Instructions defining subagent’s behavior and expertise

MCP Terms

User-wide MCP configuration (~/.verdent/mcp.json)
External tool/service integrated via Model Context Protocol
Pre-built MCP integration for specific tools or services
Project-specific MCP configuration (.verdent/mcp.json)

Performance Terms

Routing tasks to subagents to preserve main context
Multiple subagents running simultaneously
Using line ranges and search to minimize context usage
Main agent managing multiple subagent operations

See Also