Repository navigation
Prefer proportionate prevention when corrections recur - #7
Merged
Merged
Conversation
Record the first feasible rung in planning and learning, and keep a one-off finding on the existing correction path. Co-authored-by: Dennis Geldmacher <geldmacher.dennis@outlook.de>
geldmacher
changed the title
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When corrections recur, Workflow now selects the smallest sufficiently effective prevention proportionate to effort and commissioned scope. Among comparable options, it prefers structural prevention, then CI checks, bounded guidance, and human review. A first finding may also justify in-scope prevention without automatically expanding the assignment.
The central rule lives in references/learning-work.md. Planning, learning, maintained documentation, and six behavioral scenarios follow the same rule. Learning remains explicitly commissioned and limited to supported guidance; product, tool, and verifier repairs outside the assignment remain future development work.
Validation: npm run release-check passed (92 tests, no failures or skips), including all three host packages, fixed context budgets, and Markdown links. git diff --check passed. The six scenarios were inspected against the instructions; native agent trials were not run.