dd-shared-ask

Defines shared structured questioning and commit boundary rules for dd workflow skills.

Updated Jun 11, 2026
One-click install
npx skills add https://github.com/marcocpt/trae_skills --skill dd-shared-ask-marcocpt
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dd-shared-ask
Source: https://github.com/marcocpt/trae_skills/tree/main/dd-shared-ask
Command: npx skills add https://github.com/marcocpt/trae_skills --skill dd-shared-ask-marcocpt

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Multiple dd-series workflow skills duplicate the same user-questioning and commit rules, causing inconsistency and maintenance overhead. This Skill centralizes those rules so every dd skill behaves identically when asking users for decisions and committing changes. ## Core Features & Use Cases - Structured Questioning: Enforces one-question-at-a-time prompts via AskUserQuestion in Trae or request_user_input in Codex. - Null Input Re-asking: Treats null, empty, or cancelled answers as invalid and re-asks until a valid decision is obtained, never assuming defaults. - Documentation Rule Priority: Requires reading project docs like docs/CODING_STANDARDS.md before writing specs, plans, or checks. - Commit Boundaries: Restricts each commit to stage-relevant files, forbidding unrelated dirty files, secrets, --no-verify, and force pushes. - Use Case: When the dd-bug-fix-workflow skill needs user confirmation on a fix approach, it references this skill to ask a single structured question and re-ask if the user cancels. ## Quick Start Reference this skill from any dd workflow skill by adding the line: 询问规则遵循 dd-shared-ask, then follow its questioning and commit rules during execution.

Frequently Asked Questions about dd-shared-ask

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I structure user questions in an AI coding workflow?

Ask one question per AskUserQuestion call in Trae, or use request_user_input in Codex with clear options. Wait for each answer before asking the next question, never batching multiple decisions into one prompt.

What should an AI agent do when user input returns null?

Treat null, empty strings, cancellations, and unselected options as invalid input. Re-ask the original question repeatedly until a valid decision is received, and never assume a default value to continue.

How do multiple workflow skills share the same prompting rules?

Extract the common rules into a single shared skill and reference it from each workflow skill using a link like dd-shared-ask/SKILL.md. This replaces duplicated rule text and keeps behavior consistent across skills.

What commit practices should AI coding agents follow?

Each commit should contain only files relevant to the current stage. Agents must not stage unrelated dirty files, commit secrets, use --no-verify to skip hooks, or force push to remote branches.

When should project documentation rules override skill instructions?

When files like .trae/rules/docs.md, docs/CODING_STANDARDS.md, or docs/AI/trae-xctest-rules.md exist, read and follow them before writing design specs, plans, or checks. Project documentation takes priority over generic skill behavior.