From dev-workflow-skills
Use when E2E execution, failure analysis, fix-plan review, or final delivery reporting is needed before human acceptance, including E2E执行/失败分析/修复方案评审
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev-workflow-skills:delivery-qaThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This skill owns workflow Steps 19-24 and Step 26 in the delivery-stage loop after E2E assets exist. It runs E2E checks, records evidence, classifies failures with evidence, enforces the two review gates, routes fixes to the right upstream owner, and publishes the final delivery summary for human acceptance.
This skill owns workflow Steps 19-24 and Step 26 in the delivery-stage loop after E2E assets exist. It runs E2E checks, records evidence, classifies failures with evidence, enforces the two review gates, routes fixes to the right upstream owner, and publishes the final delivery summary for human acceptance.
Use these support assets during delivery work:
references/output-artifacts.mdreferences/root-cause-and-routing.mdreferences/review-gates.mdreferences/boundary-examples.mdreferences/pressure-test-prompts.md (for skill audits or revisions)Read skills/workflow-protocol/SKILL.md, then read repository-root workflow-project.yaml, then read the progress dashboard at paths.progress before touching stage work. If the request is close to a coding, test-authoring, or upstream-boundary decision, scan references/boundary-examples.md before taking ownership.
Use this skill when the human wants to:
Do not use this skill to modify production code, rewrite architecture or design, or author the E2E suite itself.
paths.acceptance<from paths.releases_dir>/r{n}/design/detail.mdpaths.architecture<from paths.releases_dir>/r{n}/design/plan.md<from paths.releases_dir>/r{n}/testing/results/run-{date}.md<from paths.releases_dir>/r{n}/testing/bug-analysis/bug-{nn}.md<from paths.releases_dir>/r{n}/testing/reviews/result-review-{nn}.md<from paths.releases_dir>/r{n}/testing/reviews/fix-review-{nn}.md<from paths.releases_dir>/r{n}/testing/results/final-delivery.mdreferences/output-artifacts.md.references/root-cause-and-routing.md and write bug-{nn}.md.references/review-gates.md and publish the review report. Stop here — do not revise yet.final-delivery.md as the cross-run delivery summary.systematic-debugger for deep evidence-first diagnosis before classifying. Delivery QA owns classification and routing; systematic-debugger owns the hypothesis-and-experiment loop.workflow-protocol.final-delivery.md summarizes the delivery outcome across runs; per-run run-{date}.md files remain the detailed records.<from paths.releases_dir>/r{n}/testing/results/run-{date}.md.bug-{nn}.md analysis with evidence-backed classification and routing.<from paths.releases_dir>/r{n}/testing/results/final-delivery.md exists once all E2E checks pass.workflow-protocol (dashboard + history) when this stage reaches its latest verified state.npx claudepluginhub yxzyzh08/dev-workflow-skills --plugin dev-workflow-skillsGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.