Skip to main content

spec-execution

6-phase iterative specification execution workflow covering implementation, testing, review, improvement, commit, and progress tracking with quality-gated convergence.

ソース情報

リポジトリ
a5c-ai/babysitter
ソースの最終更新活動
2026年6月1日 07:46
検出された SKILL.md の言語
英語
スター
1,813
フォーク
111

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

ファイルエクスプローラー
2 ファイル

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
spec-execution
description
6-phase iterative specification execution workflow covering implementation, testing, review, improvement, commit, and progress tracking with quality-gated convergence.
allowed-tools
Read, Bash, Grep, Glob
graph
{"domains":["domain:software-engineering"],"skillAreas":["skill-area:agentic-loops","skill-area:orchestration-loop"],"workflows":["workflow:feature-development"],"topics":["topic:developer-experience"],"roles":["role:tech-lead","role:backend-engineer"]}
- Read specification for requirements and architecture - Implement each requirement following project conventions - Add type definitions and JSDoc documentation - Implement error handling for all failure modes ### Phase 2: Test Writing - Write unit tests for each module/function - Write integration tests for API boundaries - Cover all acceptance criteria from the specification - Test edge cases and error paths ### Phase 3: Code Review - Verify all requirements are implemented - Check acceptance criteria are tested - Review architecture adherence - Assess code quality ### Phase 4: Iterative Improvement - Address review findings - Fix failing tests - Resolve architecture deviations - Re-verify quality (convergence loop, max 3 cycles) ### Phase 5: Atomic Commit - Group changes into logical atomic commits - Descriptive messages following project conventions - Separate production, test, and config changes - Include specification reference ### Phase 6: Progress Tracking - Requirement completion percentage - Test coverage summary - Quality score report - Remaining work identification ## Quality Gate Quality threshold defaults to 80. If not met after max improvement cycles, a human breakpoint is triggered for review. ## When to Use - `/spec:execute [file]` slash command - After spec creation when ready to implement ## Processes Used By - `claudekit-spec-workflow` (execute mode)
GitHubで見る