ワンクリックで
ll-experience-patch-settle
ADR-049 governed skill for settling validated Minor Experience Patches with backwrite RECORDS for human review.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
ADR-049 governed skill for settling validated Minor Experience Patches with backwrite RECORDS for human review.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
ADR-047 governed skill for initializing an API coverage manifest (api-coverage-manifest.yaml) from an api-test-plan, with capability × dimension expansion and four-dimensional status fields.
ADR-047 governed skill for generating structured API test specs from coverage items in api-coverage-manifest.yaml, with evidence_required, anti_false_pass_checks, and cleanup requirements.
ADR-047 governed skill for initializing an E2E coverage manifest (e2e-coverage-manifest.yaml) from an e2e-journey-plan, with four-dimensional status fields and journey-based coverage items.
ADR-047 governed skill for generating structured E2E journey specs from coverage items in e2e-coverage-manifest.yaml, with user steps, UI states, network events, anti_false_pass_checks, and evidence requirements.
ADR-047 governed skill for transforming one frozen FEAT into an API test plan (api-test-plan.md) with capability extraction, dimension matrix, and priority-based cut rules.
ADR-047 governed skill for deriving user journeys from a frozen prototype (or feat) and generating an E2E journey plan (e2e-journey-plan.md) with main journeys and minimum exception journeys.
| name | ll-experience-patch-settle |
| description | ADR-049 governed skill for settling validated Minor Experience Patches with backwrite RECORDS for human review. |
This skill implements the ADR-049 Minor patch settlement workflow. It accepts validated Patch YAML files (grade_level=minor) from the upstream ll-patch-capture skill, creates structured backwrite RECORDS in a backwrites/ subdirectory for human review, and updates the Patch status to "applied".
Critical clarification: Backwrite creates RECORDS (structured YAML summaries) -- NOT actual SSOT file modifications. A future --apply flag would perform actual SSOT modification, gated behind human confirmation. For current scope, only the record-writing path is implemented.
Major patches (grade_level=major) are REJECTED immediately with a message directing users to ll-frz-manage --type revise.
ssot/adr/ADR-049-引入体验修正层-Experience-Patch-Layer.md §4.4 (settlement rules)ssot/adr/ADR-050-SSOT语义治理总纲.md §6.2 (Minor settle behavior)ll-patch-capture (validated Patch YAML with grade_level=minor, status=approved)ssot/experience-patches/{feat_ref}/backwrites/ subdirectoryssot/adr/ADR-038-运行时核心抽象边界与对象分层基线.MD.Skill for Minor Experience Patch Settlement.ll.contract.yamlinput/contract.yamloutput/contract.yamlagents/executor.mdscripts/settle_runtime.pystatus: approved and grade_level: minor.status is already "applied", return early (no-op, not an error).change_class in BACKWRITE_MAP to get the list of backwrite targets.ssot/experience-patches/{feat_ref}/backwrites/{target}_updates.yaml. These are RECORDS for human review, NOT actual SSOT file modifications.status: applied and settled_at: <ISO timestamp>.| change_class | must_backwrite_ssot | backwrite_targets |
|---|---|---|
| ui_flow | False | ui_spec_optional |
| copy_text | False | (none) |
| layout | False | ui_spec_optional |
| navigation | True | ui_spec, flow_spec |
| interaction | True | ui_spec, flow_spec, testset |
| error_handling | False | (none) |
| performance | False | (none) |
| accessibility | False | ui_spec_optional |
| data_display | False | ui_spec_optional |
| visual | False | ui_spec_optional |
| semantic | True | frz_revise (NOT handled by settle) |
| other | False | (none) |
ll-patch-capturebackwrites/ subdirectory--apply flag), Major patch processing (routed to FRZ revise)ll-frz-manage --type revise.--apply flag would perform actual SSOT modification, gated behind human confirmation.test_impact.affected_routes is non-empty.yaml.safe_load() only -- no yaml.load() for YAML deserialization.derive_grade from cli/lib/patch_schema.py -- do NOT re-implement grade derivation locally.other fallback to handle unknown values gracefully.