一键导入
sddp-qc
Execute Quality Control checks on the implemented feature to ensure requirements, security, and tests pass
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Execute Quality Control checks on the implemented feature to ensure requirements, security, and tests pass
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Perform non-destructive cross-artifact consistency and quality analysis across spec, plan, and tasks
Run the full SDD pipeline (Specify → Clarify → Plan → Checklist → Tasks → Analyze → Implement+QC) end-to-end without user interaction
Generate a custom requirements quality checklist for the current feature
Identify underspecified areas in a feature spec and resolve them through targeted clarification questions
Create or refine the canonical project-level deployment and operations context (`specs/dod.md`)
Analyzes the repository and guides the user through setting up their local development environment
| name | sddp-qc |
| description | Execute Quality Control checks on the implemented feature to ensure requirements, security, and tests pass |
You are starting a Quality Control workflow. Your sole purpose is to verify the code written in the implementation step against specifications and quality standards. Disregard any prior specification or planning discussion from this conversation. Focus exclusively on quality control.
Load and follow the workflow in .github/skills/quality-control/SKILL.md.
When the shared workflow requires user decisions and AUTOPILOT = false:
When AUTOPILOT = true, keep following the shared workflow's automatic decision rules unchanged.
When the workflow says Delegate, read the referenced sub-agent file at that point, not before — then perform the task yourself:
.github/agents/_context-gatherer.md.github/agents/_qc-auditor.md.github/agents/_story-verifier.mdThis adapter does not declare a native browser tool. The shared QC workflow still runs the Step 6.0 browser probe against any browser-capable tools the current harness exposes. If no browser-capable tool is reachable, follow the terminal/headless and manual-test.md fallback paths.
Report progress to the user at each major milestone — summarize what has been checked and what issues were found.