Executes individual user stories and tasks autonomously within defined project boundaries and git worktrees.
The Story Executor skill transforms Claude into an autonomous task agent capable of completing atomic units of work defined in a tasks file. It manages the entire implementation lifecycle—from reading project-specific guidelines in CLAUDE.md to verifying acceptance criteria, running typechecks, and performing automated commits. Designed for scalability and security, it enforces strict project boundaries to prevent unauthorized file access and natively supports git worktrees, allowing for efficient parallel task execution across complex codebases.
主な機能
01Autonomous execution of atomic tasks in a single context window
02Git worktree integration for parallel task processing
033 GitHub stars
04Strict project root boundary enforcement for enhanced security
05Automated verification via typechecking and acceptance criteria mapping
06Standardized git commit generation and detailed failure reporting
ユースケース
01Ensuring consistent code quality by enforcing CLAUDE.md guidelines during implementation
02Implementing specific feature requirements or bug fixes from a prioritized task list
03Scaling development capacity by running multiple task agents in parallel worktrees
