AI Agent Hub
Back to plugins
🖥️

dsh-copy-tool

Client Updated 2026.09.09

Run the following command in DeepSeek Harness:

dsh plugin install kezan31/dsh-copy-tool

Paste the following prompt into your AI chat to install this plugin:

Install in DeepSeek Harness by running dsh plugin install kezan31/dsh-copy-tool, available at https://github.com/kezan31/dsh-copy-tool, then restart DSH Web to activate.

About this plugin

When refactoring large projects with the write tool, AI models frequently drop lines, oversimplify complex logic, and produce incomplete patches that cost extra tokens and rounds of修补. dsh-copy-tool turns "move these exact lines there" from a vague context reproduction into a deterministic, line-number-driven operation: you specify which lines go where, and it migrates them verbatim—no more, no less.\n\nUnder the hood, a single call can mix discrete lines and continuous ranges (e.g. [633, [640, 645], 700]) with source lines and target lines mapped positionally; it detects and strips the common indentation across all selected non-empty lines, then re-applies a clean four-space indent level via the indent parameter. A dry_run pass emits a bounded unified diff for review before anything touches the filesystem. Creating a missing file and editing an existing one (including the source file itself) are both atomic operations that preserve untouched content and filesystem-native line endings, with every path resolved relative to the current session workspace and governed by its sandbox policy.\n\nIf you are splitting modules, relocating functions across files, or batch-adjusting indent levels inside DSH—or simply want your model to spend tokens on logical decisions instead of re-typing code—this plugin makes line-level copying a previewable, auditable, and reversible operation that significantly cuts the rework caused by omissions and oversimplification.

Use Cases

  • Migrate functions across files by specifying exact source and target lines to avoid AI omissions or rewrites
  • Batch-adjust indent levels with automatic common-indent stripping and clean re-alignment
  • Preview a bounded unified diff via dry_run before atomic writes to minimize rework

Best For

  • Developers refactoring large projects in DeepSeek Harness who frequently move code across files
  • Engineers who want to cut token spend and avoid AI oversimplifying complex logic
  • Developers who need precise indent control and line-level editing with previewable, reversible writes