
Call Cursor Agent
- 413 installs
- 3 repo stars
- Updated January 23, 2026
- dotneet/claude-code-marketplace
call-cursor-agent is a Claude Code skill that delegates coding subtasks to the cursor-agent CLI and merges the returned work back into the same agent workflow.
About
call-cursor-agent is a cross-agent integration skill for developers running Claude Code who want Cursor to execute a bounded task and return results into the same session. The skill checks for cursor-agent with which, then invokes cursor-agent --model with a task heredoc, defaulting to composer-1 when no model is specified. It fits parallelizing implementation, exploratory edits, or offloading a slice while Claude Code retains orchestration. Use call-cursor-agent when cursor-agent is installed and a subtask is concrete enough to hand off via CLI.
- Cross-agent delegation
- Cursor agent invocation
- Workflow handoff patterns
- Parallel coding automation
Call Cursor Agent by the numbers
- 413 all-time installs (skills.sh)
- Ranked #1,942 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Jul 29, 2026 (Skillselion catalog sync)
npx skills add https://github.com/dotneet/claude-code-marketplace --skill call-cursor-agentAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 413 |
|---|---|
| repo stars | ★ 3 |
| Last updated | January 23, 2026 |
| Repository | dotneet/claude-code-marketplace ↗ |
How do you delegate tasks from Claude Code to Cursor?
Delegate coding subtasks from Claude Code to a Cursor agent and merge results back into the same workflow.
Who is it for?
Developers using Claude Code with cursor-agent installed who want Cursor to execute scoped coding tasks via CLI delegation.
Skip if: Environments without cursor-agent installed or workflows that require a single agent with no cross-tool handoff.
When should I use this skill?
Claude Code should hand a concrete coding subtask to cursor-agent and integrate the CLI response.
What you get
cursor-agent CLI output, completed subtask edits, and merged results in the Claude Code workflow.
- cursor-agent task output
- merged code changes
By the numbers
- Defaults to composer-1 when cursor-agent model is unspecified
Files
Call cursor-agent to perform a task
This skill requests a task to cursor-agent.
Checking for the Existence of cursor-agent
# Check if cursor-agent is installed. Exit code 0 means success, 1 means failure.
# If not installed, skip the subsequent steps of the skill.
which cursor-agent
if [ $? -ne 0 ]; then
echo "cursor-agent is not installed"
fiRequesting a Task to cursor-agent
# If the model is not specified, the default model is composer-1.
model=composer-1
cursor-agent --model=$model <<EOT
{task_description}
EOTRelated skills
FAQ
What is the default model for call-cursor-agent?
call-cursor-agent defaults to composer-1 when no model is specified, passing the task to cursor-agent via a shell heredoc after verifying cursor-agent is installed with which.
What happens if cursor-agent is not installed?
call-cursor-agent runs which cursor-agent first and stops subsequent steps when the CLI is missing, preventing failed delegation attempts inside Claude Code.