Add prompt-fieldtest skill - #1680
Open
Dragusin-Cristian wants to merge 3 commits into
Open
Conversation
Field-tests an instruction document (a rule file, CLAUDE.md, an agent, another skill) by having a blinded builder agent implement a throwaway feature guided only by that document, a read-only auditor agent review the result clause by clause, and a scoped cleanup agent erase the probe. Registers the skill as its own plugin in the marketplace, matching the other single-skill plugins. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Matches the LICENSE.txt and `license:` frontmatter key carried by the other open-source skills in this repo. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Reading an instruction document tells you whether it's agreeable, not whether it's followable. Normally you find that out by using the rule for a week and watching the agent drift — about as expensive as writing the feature yourself.
This skill does it in one run. A fresh agent builds a throwaway feature guided only by the document, a second audits what the document caused, a third erases the probe. The deliverable is a report about the document; the code is evidence, and it always dies.
Five sequential phases: design a brief that fires most of the document's clauses → build blind → audit without the builder's self-report → verify the agents' claims against the real toolchain → path-scoped cleanup against a baseline captured up front.