How this command is triggered — by the user, by Claude, or both
Slash command
/cursor-opinion:delegate [task description]This command is limited to the following tools:
The summary Claude sees in its command listing — used to decide when to auto-load this command
# Delegate Task to Cursor The user wants Cursor AI to make code changes in the current project. ## User's Task Description $ARGUMENTS ## Important: This runs Cursor with WRITE permissions Cursor will be given full access to read and modify files in the current working directory. This is a SLOW operation - expect 2-5 minutes. When calling Bash, you MUST set a long timeout: - Use `timeout: 600000` (10 minutes) for the Bash tool call ## Your Task 1. **Identify the Task**: Determine what code changes the user wants made. - If a task description was provided above, use that as the pri...
The user wants Cursor AI to make code changes in the current project.
$ARGUMENTS
Cursor will be given full access to read and modify files in the current working directory. This is a SLOW operation - expect 2-5 minutes.
When calling Bash, you MUST set a long timeout:
timeout: 600000 (10 minutes) for the Bash tool callIdentify the Task: Determine what code changes the user wants made.
Gather Context: Collect relevant context including:
Build a Comprehensive Prompt: Create a detailed, actionable prompt that tells Cursor exactly what changes to make. Include:
Execute Delegation: Call the delegate script with the working directory.
CRITICAL: Set timeout to 600000ms (10 minutes) because Cursor uses a slow model:
# timeout: 600000
"${CLAUDE_PLUGIN_ROOT}/scripts/delegate-cursor.sh" "YOUR_COMPREHENSIVE_PROMPT" "$(pwd)"
Report Results: When Cursor completes, summarize:
$(pwd))agent command is not found, inform the user they need to install the Cursor CLInpx claudepluginhub glekner/cursor-opinion-marketplace --plugin cursor-opinion/executeDelegates execution of a specific plan step to Cursor (Agent mode, Auto model) for file operations, then runs verification commands. Supports --background, --resume, --model flags.
/codexDelegates coding tasks to OpenAI Codex CLI for autonomous execution via natural language prompts, handling code refactoring, testing, bug fixes, explanations, features, and reviews.
/delegateGenerates Codex prompt for code/experiment tasks from project context files, confirms plan with user, displays new-terminal execution command, polls agent_handoff.md for results, and prompts next steps.
/delegateOrchestrates complex tasks: decomposes into agent waves/tasks, plans execution, awaits user approval, then executes via subagents or teams.
/delegateDispatches implementation tasks to subagents using a plan, then auto-chains to review. Also supports --fixes and --pr-fixes modes for addressing issues.
/delegateDelegates a task to a named altana harness preset, runs it with optional prompt template and write access, then surfaces a structured JSON result with status, duration, response, and log path.