mirror of
https://github.com/QwenLM/qwen-code.git
synced 2026-04-28 03:30:40 +00:00
* feat(skills): add /batch skill for parallel batch operations Add a new built-in skill `/batch` that orchestrates large-scale parallel changes across multiple files. The skill automatically: - Discovers target files using glob patterns - Splits files into chunks for parallel processing - Launches multiple worker agents concurrently - Aggregates results with success/failure statistics - Supports --dry-run mode for preview Resolves #3043 * fix(skills): address PR review feedback for /batch skill - Extend chunking table to cover 51-100 files (now uses 5 chunks) - Clarify that `task` tool is the Agent tool for spawning workers - Add SKIPPED status to Agent Prompt Template output format - Shorten description field, move examples to body - Expand test file exclusion patterns (*.test.js, __tests__/, etc.) - Expand Dry-Run Mode section with detailed example output - Fix example math: "24 files → 3 chunks of ~8 each" * fix(skills): address wenshao's review feedback for /batch skill - Add copyright header - Add ask_user_question to allowedTools - Fix edit_file -> edit (canonical tool name) - Fix chunking table math (3-8 each, not ~7-8) - Add zero file match handling - Reframe dry-run from --flag to natural language pattern * fix(skills): remove unnecessary copyright header from SKILL.md --------- Co-authored-by: jinye.djy <jinye.djy@alibaba-inc.com> |
||
|---|---|---|
| .. | ||
| channels | ||
| cli | ||
| core | ||
| sdk-java | ||
| sdk-typescript | ||
| vscode-ide-companion | ||
| web-templates | ||
| webui | ||
| zed-extension | ||