Deep Executor Skill
Activate autonomous deep work mode for complex tasks.
Overview
Deep Executor is the autonomous deep worker agent. When activated, it:
-
Explores First: Uses its own tools (Glob, Grep, Read, ast_grep_search) to thoroughly understand the problem
-
Plans Strategically: Creates execution plan based on exploration
-
Executes Directly: Does all work itself using Edit, Write, Bash, ast_grep_replace
-
Verifies Everything: Runs builds, tests, and diagnostics before claiming completion
-
Completes 100%: Guarantees full completion with evidence
Usage
/oh-my-claudecode:deep-executor <your complex task>
Or use magic keywords:
-
"deep-executor: ..."
-
"deep work: ..."
-
"forge: ..."
When to Use
Situation Use Deep Executor?
Complex multi-file refactoring YES
Unclear implementation path YES
Need guaranteed completion YES
Simple single-file fix NO (use executor)
Quick code search NO (use explore)
Cost-sensitive work NO (use ecomode)
Activation
This skill spawns the deep-executor agent via:
Task(subagent_type="oh-my-claudecode:deep-executor", model="opus", prompt="{{PROMPT}}")
The agent handles all exploration and execution internally. No sub-agents are spawned.