refine: strengthen failure-mode and spec-completeness coverage in planning agents#179
Merged
shinpr merged 3 commits intoJun 23, 2026
Merged
Conversation
…coverage Close gaps where a feature's combinatorial shape (interacting modes, ordering of absent values, reset/clear lifecycle) could pass through planning, decomposition, and review without becoming a verifiable contract. - Failure Mode Checklist: add ninth domain-independent category `missing-sort-key ordering` (work-planner, document-reviewer, plan-template) - AC drafting: add a mode x branch combination expansion step (technical-designer, technical-designer-frontend) - State lifecycle: capture reset/clear as a state-lifecycle-negative contract — Reset/Clear Behavior column in the Client State Design table and a work-planner rule to record it in Reference Contract Values - Failure Mode propagation: propagate applicable categories to task Proof Obligations (task-decomposer) and verify them in review (code-reviewer, integration-test-reviewer, task-template), with a deterministic fallback when taskFiles are absent so AC-less obligations are not silently treated as verified Scope: English (-en) agent/skill sources only. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…age to ja Japanese (-ja) counterpart of the prior -en change, keeping wording, tone, and terminology consistent across both locales. - Failure Mode Checklist: add ninth domain-independent category `missing-sort-key ordering` (work-planner, document-reviewer, plan-template) - AC drafting: add a mode x branch combination expansion step (technical-designer, technical-designer-frontend) - State lifecycle: capture reset/clear as a state-lifecycle-negative contract — reset/clear behavior column in the Client State Design table and a work-planner rule to record it in Reference Contract Values - Failure Mode propagation: propagate applicable categories to task Proof Obligations (task-decomposer) and verify them in review (code-reviewer, integration-test-reviewer, task-template), with a deterministic fallback when taskFiles are absent so AC-less obligations are not silently treated as verified Technical terms and tokens (e.g. state-lifecycle-negative, Proof Obligations, coverage_gap, needs_improvement, taskFiles) kept in English; prose mapped to the existing ja vocabulary. Scope: Japanese (-ja) agent/skill sources only. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Bump patch version and record the failure-mode / spec-completeness coverage changes in the changelog. Co-Authored-By: Claude Opus 4.8 (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.
Summary
Close gaps where a feature's combinatorial shape — interacting modes, ordering of absent values, reset/clear lifecycle — could pass through planning, task decomposition, and review without ever becoming a verifiable contract. The changes anchor to existing mechanisms (Failure Mode Checklist, AC drafting, Reference Contract Values, Proof Obligations) rather than adding parallel structures, and are applied identically across en and ja.
Changes
missing-sort-key ordering(ordering when some elements lack the sort key). Propagated to the enumerations inwork-planner, thedocument-reviewercompleteness gate, and the plan template.mode × branchcombination-expansion step intechnical-designerandtechnical-designer-frontend: when a change adds a mode/flag/variant overlaying an existing selection/ordering/filtering/display branch, expand the new value against each existing branch (a mode can take effect on one branch while silently no-opping on the others).state-lifecycle-negativecontract: a Reset/Clear Behavior column in the Design Doc Client State Design table and awork-plannerrule to record the reset-returns-to-unused condition in Reference Contract Values.task-decomposer), and verify them in review (code-reviewer,integration-test-reviewer, task template) — including a deterministic fallback whentaskFilesare absent, so AC-less obligations are not silently treated as verified.Scope & validation
-enand-jaagent/skill sources; per-file diff shape is identical across locales.1.24.2with a changelog entry.check:skills-indexpre-commit hook passes (26 skills consistent across en/ja).🤖 Generated with Claude Code