refine(design-doc): evidence-backed behavioral claim verification - #183
Merged
Conversation
… (en) Add an assumption check so behavioral/factual claims a design relies on (framework defaults, capabilities assumed provided, features assumed implemented) are each backed by evidence at design time or bound to a follow-up, running end-to-end across generator, template, and reviewer. - technical-designer(+frontend): new "Behavioral Claim Verification" step under Existing Code Investigation. Materiality-scoped (claims the design does not itself define, whose falsity invalidates the approach). Evidence or Confirmed: No + matching Risks and Mitigation row propagated downstream as `verify at [step or artifact]`. Framework defaults require doc paired with the resolved package version. Routing/gate-timing vs Fact Disposition Table and Cross-Layer Assumptions made explicit. - design-template: "Assumed Behaviors" slot in the Agreement Checklist, reusing the Evidence/Confirmed pattern and keyed to Risks and Mitigation. - document-reviewer: behavioral-claim evidence check in the quality gate (positive-form exclusion for correctly-routed claims; flags missing slot, unbacked Confirmed: Yes, missing version pairing, or missing downstream propagation). Also compress redundancy without changing requirements: collapse tautological skill cues, condense the Change Impact example to the three impact tiers, and trim Fact Disposition prose duplicated by the generator. Restore the formal Serialized Format / Consumer Parse Rule terms in the template to keep the downstream contract vocabulary aligned. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Port the evidence-backed behavioral claim verification into the Japanese sources so ja matches en end-to-end (generator, template, reviewer). - technical-designer(+frontend): new "振る舞いに関する主張の検証" step under Existing Code Investigation, plus the same compressions (skill-cue collapse, Change Impact condensed to three tiers). - design-template: "Assumed Behaviors" slot in the Agreement Checklist, plus the Fact Disposition prose trim and the Serialized Format / Consumer Parse Rule alignment. - document-reviewer: behavioral-claim evidence check in the quality gate. Localization choices: keep coined mechanism names (Assumed Behaviors, Fact Disposition Table, Cross-Layer Assumptions) and framework terms in English; reuse existing ja renderings (確認: 済/未, 根拠, 検証先: [ステップまたは成果物], リスクと対策, 作業計画書) to avoid notation drift; add Japanese scan-start cues (既に / デフォルトで / 処理済み …) so the reviewer catches Japanese claims; "再掲" (not 言い換え) preserves the shared lookup key. check:skills-index passes; en/ja consistent across all 28 skills. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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
Design Docs frequently rely on unstated behavioral or factual claims — a framework/library default, a capability assumed already provided, or a feature assumed already implemented — that then flow into implementation unverified. The existing dependency check only confirms a component exists, not that it behaves as the design assumes.
This adds an evidence-backed assumption check that runs end-to-end across the generator, the template, and the reviewer, so each relied-upon behavioral claim is either backed by evidence at design time or explicitly bound to a follow-up.
What changed
technical-designer, backend + frontend) — a Behavioral Claim Verification step (Gate 1) for each material claim the design relies on but does not itself define. Evidence-backed → recorded in the Agreement Checklist "Assumed Behaviors" slot withConfirmed: Yes; otherwiseConfirmed: No+ a matching Risks and Mitigation row propagated downstream asverify at [step or artifact]. A framework/library default requires the official doc paired with the resolved package version. Routing defers to the Fact Disposition Table (Codebase Analysis facts) and Cross-Layer Assumptions (prior-layer claims).design-template) — an Assumed Behaviors slot in the Agreement Checklist, reusing the existing Evidence / Confirmed pattern and keyed to Risks and Mitigation.document-reviewer) — a behavioral-claim evidence check in the quality gate (important/feasibility), with a positive-form exclusion for claims correctly routed elsewhere.Also compresses redundancy without changing requirements (skill-cue collapse, Change Impact condensed to three tiers, duplicated Fact Disposition prose trimmed).
Scope
Applied across en/ja sources;
check:skills-indexconsistent across all 28 skills. Patch bump to1.25.2.🤖 Generated with Claude Code