claude-code-x_skill

Overview

This skill runs Claude Code autonomously in the background to complete well-defined development tasks. It supports up to 5 concurrent sessions, non-blocking operation, and persistent JSON-based session tracking so you can continue other work while tasks run.

How this skill works

Validate each task for clarity and independence, then spawn a background Claude Code session via the provided shell script. Sessions execute with configured tools (Bash, Read, Write, Edit, Glob, Grep), log output to disk, and expose commands to check status, retrieve results, kill sessions, or clean up. Session metadata and logs are persisted for monitoring and post-run reporting.

When to use it

• You have 3+ independent, well-scoped code tasks that don’t need interactive decisions.
• You want to run similar tasks across multiple modules in parallel (e.g., add logging to several crates).
• You need non-blocking automation so you can continue other work while tasks run.
• You want to maximize throughput with up to 5 concurrent workers.
• You need repeatable, scriptable background execution with retrievable logs and session IDs.

Best practices

• Define tasks with clear success criteria and required file targets before launching.
• Group similar tasks for full parallel efficiency; use phased hybrid approaches for mixed complexity.
• Set an appropriate timeout per task; increase default if tasks are larger than 10 minutes.
• Capture and store session IDs immediately to enable monitoring and result retrieval.
• Check logs on failure to decide if a retry or manual fix is needed; include missing-file creation in task text when necessary.

Example use cases

• Add comprehensive logging to five core modules by launching five parallel sessions and monitoring completion.
• Refactor error handling in a single crate with a custom timeout and toolset, then retrieve the result log.
• Run infrastructure setup tasks in parallel, then switch to interactive sessions for security-critical flows.
• Automate repetitive edits across multiple files (e.g., update imports or apply consistent tracing macros).
• Spawn background sessions to run tests or lint changes while continuing other development work.