code-agentic

What problem does it solve?

AI-driven code execution, especially for destructive or high-risk operations, requires robust safety protocols to prevent errors, data loss, and token waste. This skill defines mandatory verification gates, rollback mechanisms, and confidence thresholds for all Claude Code operations.

Core Features & Use Cases

• Dependency Exclusion Enforcement: Critically prevents reading from large dependency/build directories (e.g., node_modules/), saving massive token costs.
• Pre-Execution Verification: Mandates documenting assumptions, checking permissions, dependencies, environment, and Git state before any high-risk operation.
• Error Recovery & Rollback: Establishes clear rollback plans, including Git safety branches and stash backups, for every operation.
• Use Case: You're about to perform a database migration, install a new package, or modify critical files, and you need to ensure the operation is safe, reversible, and fully verified.

Quick Start

"Perform a pre-execution verification for installing a new npm package." "Generate a rollback plan for a database schema migration."