feat: add subagent-driven-prompt-tuning skill#26
Merged
Conversation
…ctice Add a new skill that tunes Claude Code prompts (SKILL.md, slash commands, task prompts, CLAUDE.md sections) by dispatching unbiased subagents to execute them, then iterating on both executor self-report and caller-side metrics until improvements plateau. Adapted from mizchi/chezmoi-dotfiles's empirical-prompt-tuning skill with attribution, renamed to better signal what is being tuned and how.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a new
subagent-driven-prompt-tuningskill to theclaude-code-best-practiceplugin.The skill tunes Claude Code prompts (SKILL.md, slash commands, task prompts, CLAUDE.md sections, code-gen prompts) by dispatching an unbiased subagent to execute them, then iterating based on:
[critical]requirements, accuracy %,tool_uses,duration_msIt halts only when two consecutive iterations show no new ambiguity and metric deltas fall below threshold, with an overfitting check on a hold-out scenario.
Changes
plugins/claude-code-best-practice/skills/subagent-driven-prompt-tuning/SKILL.mdCLAUDE.mdplugin table to list all five skills underclaude-code-best-practiceAttribution
Adapted from
empirical-prompt-tuningin mizchi/chezmoi-dotfiles, rewritten to match this marketplace's tone and renamed to signal what is being tuned and how. Attribution is preserved in the SKILL.md body.Test plan
claude-code-best-practicewhen the plugin is installed