Repository navigation
feat(review): read the implementer report as self-grading and keep two honesty lists - #11
Merged
Merged
Conversation
…two honesty lists The Code Reviewer took the implementer's report at face value and had no place to record what it set aside or could not check, so a rationale in the report could talk a finding down and an unverifiable check was either raised on suspicion or dropped. The reviewer now grades the package, treats pasted output as evidence of the command it shows and nothing more, never re-runs a test the report evidences, grades a correctness defect by its effect on a reasonable user even when no criterion names it, and returns two lists the code-review schema now carries: declined_to_judge for every behavior set aside with its reason, and cannot_verify for every check the package cannot answer with the command that would, which Check runs. The reviewer prompt never caps a severity or excludes an area, and the reviewer ignores such a sentence and notes it. On a fix dispatch a developer agent verifies each finding against the code before applying it and says so with evidence when one is wrong. Source of the pattern: obra/superpowers requesting-code-review and subagent-driven-development reviewer prompts, https://github.com/obra/superpowers Co-Authored-By: Claude Fable 5.1 <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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Area 3 of 11: read the implementer report as self-grading, keep two honesty lists
The Code Reviewer grades the package, treats pasted output as evidence of the command it shows, never re-runs an evidenced test, grades a correctness defect by its effect on a reasonable user even when no criterion names it, and returns two lists the code-review schema now carries:
declined_to_judgeandcannot_verify(with the command that would answer it, which Check runs). The reviewer prompt never caps a severity or excludes an area. On a fix dispatch a developer agent verifies each finding against the code before applying it.Source: obra/superpowers reviewer prompts.
Stacked PR: merge in order, this one after its base branch. Merge with a merge commit, as the earlier PRs were.
🤖 Generated with Claude Code