Task Clarifier: When Your Agent Needs to Ask Better Questions

Where this fits

You're building an agent that books meetings. A user types "set up a review with John next week." Without clarification, your agent guesses the day, time, duration, and agenda — and likely gets it wrong. The task-clarifier skill solves this by forcing a structured interrogation loop before any action. It's designed for developers who want their agents to reduce ambiguity and avoid costly rework, especially when user requests are vague or under-specified.

Why agents benefit

• Forces explicit confirmation — The agent won't proceed until the user confirms the AI's understanding is correct, eliminating silent assumptions.
• Uncovers latent contradictions — It probes for hidden constraints (e.g., "next week" might conflict with a holiday) that the user didn't mention.
• Adapts to user language automatically — All questions and summaries are generated in the user's own language, defaulting to Chinese when unknown, which improves accessibility.
• Prevents scope creep — By clarifying the true need, the agent avoids doing extra work on dimensions the user doesn't care about.
• Turns vague requests into actionable specs — The three-goal framework ensures the AI has a complete, unambiguous picture before executing.

Practical scenarios

Scheduling and Calendar Management

A user says "set up a review" without specifying participants, format, or location. The skill asks about time zones, preferred tools (Zoom vs. in-person), and whether a written summary is needed — turning a one-liner into a fully specified event.

Code Generation and Refactoring

A developer asks "refactor this module" without stating goals. The skill probes for priorities: performance, readability, test coverage, or minimal diff. It also checks for hidden dependencies or legacy constraints, ensuring the refactor aligns with actual needs.

Data Analysis and Reporting

A stakeholder requests "insights from sales data" but hasn't defined metrics or time range. The skill clarifies the business question, the granularity (daily vs. monthly), and the output format — preventing a generic dashboard that misses the point.

Add it to your agent workflow

Install the skill via:

npx skills add dongshuyan/compass-skills --skill task-clarifier

Then invoke it in your agent's prompt when you detect ambiguity. For example, before executing a user request, add a step:

If the user's request is unclear or incomplete, call $task-clarifier to clarify their needs before proceeding.

The skill will handle the questioning loop and only return when all three goals are met.

Compare before adopting

Before integrating, evaluate the skill against your needs:

• Quality signals — Check the repo's commit history, issue resolution, and whether the skill handles edge cases like multi-turn conversations.
• Maintenance freshness — Look at last commit date and community activity; a stale skill may not keep up with LLM behavior changes.
• Alternatives — Compare with built-in prompting patterns or other clarification skills. Some may be lighter-weight if you only need a single confirmation.
• Workflow fit — Consider if the strict three-goal loop might over-clarify trivial requests, adding friction for power users.

Why it is worth tracking

With 721 stars and an active author, this skill shows strong community interest and a clear design philosophy. Its strict requirement to confirm understanding before acting is a quality signal that resonates with developers building robust agents. If you're handling ambiguous user input, this skill is worth evaluating now. For more details, see the skill page at https://www.openagentskill.com/skills/dongshuyan-task-clarifier.