batch-edit

Applies uniform transformations to many file entries using parallel subagents on split chunks.

1|Updated Dec 18, 2009
One-click install
npx skills add https://github.com/thurn/dotfiles --skill batch-edit-thurn
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: batch-edit
Source: https://github.com/thurn/dotfiles/tree/main/.llms/skills/batch-edit
Command: npx skills add https://github.com/thurn/dotfiles --skill batch-edit-thurn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Applying the same edit to hundreds of entries in a large file is slow and error-prone when done sequentially, and letting multiple agents edit one file causes conflicts. This Skill splits large files into entry-aligned chunks, processes them with parallel subagents, and reassembles the results safely. ## Core Features & Use Cases - Chunked Parallel Processing: Splits files at entry boundaries (TOML arrays, Markdown sections, JSON elements) and processes chunks with up to 5 concurrent subagents. - Two I/O Modes: Transform mode edits chunks in place; generate mode reads input chunks and writes results to separate output files. - Autonomous Parameter Inference: Detects delimiters, batch sizes, and model choice (Sonnet vs. Opus) without asking the user. - Use Case: Rewrite a field in every entry of a 500-entry TOML config file, or generate descriptions for every item in a large Markdown catalog, in a fraction of the sequential time. ## Quick Start Apply this transformation to every entry in my large config file using parallel batch editing.

Frequently Asked Questions about batch-edit

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I apply the same edit to many entries in a large file?▼

Split the file into chunks at entry boundaries, then process chunks in parallel with subagents that each edit their own chunk file. After all rounds complete, concatenate the chunks back into the original file and verify entry counts.

How to run parallel file edits without agent conflicts?▼

Never let two subagents edit the same file. Write each chunk to its own file in a temporary directory, assign one chunk per subagent, and reassemble only after all rounds finish.

What file formats work with batch chunk splitting?▼

Any format with repeating entry structures works: TOML [[array]] headers, Markdown ## sections, blank-line-separated blocks, XML tags, and JSON array elements. The delimiter is inferred by inspecting the first ~100 lines.

When should I use transform mode vs generate mode?▼

Use transform mode when output replaces the input, such as rewriting fields in place. Use generate mode when input and output differ, such as reading source entries and writing generated content to a separate destination file.

What happens when a subagent fails or skips entries?▼

Spot-check one chunk per round by reading a few entries. If a subagent failed or skipped entries, re-run that specific chunk in a later round rather than skipping it.