一键导入
dev-workflow
Start working on a ticket, dev workflow, work on an issue, pick up a ticket
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
Start working on a ticket, dev workflow, work on an issue, pick up a ticket
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
Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Spec (does the code match what the originating issue/PRD asked for?). Runs both reviews in parallel sub-agents and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.
Build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model.
Set up Claude Code hooks to block dangerous git commands (push, reset --hard, clean, branch -D, etc.) before they execute. Use when user wants to prevent destructive git operations, add git safety hooks, or block git push/reset in Claude Code.
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
| name | dev-workflow |
| description | Start working on a ticket, dev workflow, work on an issue, pick up a ticket |
| disable-model-invocation | false |
| argument-hint | [issue-number-or-description] |
Run the full developer workflow end-to-end: pick a ticket, branch, plan, implement, commit, simplify, PR, wait for CI, and address feedback. Execute each step sequentially, waiting for completion before moving to the next.
Ticket selection:
$ARGUMENTS is provided, treat it as an issue number or description.gh issue list --assignee @me --limit 10 to show issues assigned to the user.gh issue list --limit 10.Ticket creation (if no existing ticket matches):
/tmp/issue-draft.md).gh issue create with the approved values.Branch creation: Invoke the cut-branch skill with the issue number.
Planning: Invoke EnterPlanMode (the built-in /plan command). Present a plan for implementing the ticket. Wait for the user to approve the plan before continuing.
Implementation: Execute the approved plan. Build the feature or fix step-by-step.
Visual/functional verification (conditional): If the implementation touched UI components, styles, layouts, or templates, spawn a Playwright subagent to verify the feature works as intended. The agent should use its best judgment on verification method (screenshots, checking DOM elements, verifying text content, testing interactions, etc.). If verification passes, proceed without waiting for approval. If it fails, fix the issue before continuing.
Initial commit: Invoke the commit-commands:commit skill.
Simplify: Only after step 7's commit has completed, invoke the simplify-agent skill on the changed code. Do not run commit and simplify in parallel or batch them in one tool turn — the commit must be on disk before simplify reads the tree, and simplify's edits must land as a separate follow-up commit. This step is for pre-PR cleanup — refactoring to cleaner code, removing duplication, improving reuse. It is NOT for addressing PR review feedback (that's step 11). The simplify-agent skill launches code-simplifier agents that review for reuse, quality, and efficiency issues, then fix what they find.
Push & PR: Invoke the commit-commands:commit-push-pr skill.
Wait for CI: Spawn a background agent that watches gh pr checks until all checks complete, then reports pass/fail. Continue to the next step only after the background agent completes.
Address feedback: Invoke the address-pr-feedback skill with the PR number. This handles automated review bot feedback and human reviewer comments — different from the simplify step which is a pre-PR self-review.
Final review: Ask the user if they have further feedback on the PR. If so, address it. If not, the workflow is complete.