Grilling with Letters: Better Agent Q&A Format

Where this fits

You're building an agent that 'grills' users—asking tough, Socratic questions to sharpen their thinking. But your users keep stumbling over the answer format: numbered options confuse them, or they feel forced to pick a choice when they want to explain themselves. The grilling-fix skill solves this by enforcing lettered options (A, B, C…) and always allowing free-form answers, while keeping every other grilling behavior intact.

Why agents benefit

• Consistent option presentation: Every question gets explicit letters (A, B, C…) instead of numbers, reducing parsing ambiguity in downstream logic.
• Preserves user autonomy: Users can always answer in their own words, so the agent doesn't lose valuable context from rigid multiple-choice constraints.
• Seamless integration: It overrides only the answer-option format; all other grilling instructions remain unchanged, so you don't rewrite your whole agent flow.
• Supports multi-select: The skill handles single- or multi-select options as the question demands, giving you flexibility in response design.
• Clear trigger mechanism: It activates only when explicitly invoked (e.g., /grilling /grilling-fix or a request for lettered options), preventing accidental overrides.

Practical scenarios

Interview Prep Coach

An agent drills a user on system design questions. With grilling-fix, each question offers lettered trade-off options (A, B, C…) but also invites a free-text explanation. The user can pick a letter and then elaborate, or skip the letter entirely and answer in their own words—keeping the session flexible.

Decision-Making Advisor

A career coach agent asks 'Would you rather optimize for salary or growth?' Instead of forcing a binary, it presents options labeled A and B, but the user can respond with 'I'd consider both depending on the company stage.' The agent adapts without losing the structured format.

Study Buddy Quizzer

A language-learning agent quizzes vocabulary. Lettered options (A, B, C, D) make it easy for the user to select quickly, but the skill also allows typed answers, so the agent can catch nuanced responses that predefined choices would miss.

Add it to your agent workflow

Install the skill:

npx skills add eliasstravik/skills --skill grilling-fix

Then invoke it alongside grilling in your agent:

User: /grilling /grilling-fix
Agent: Question 1: What's the main risk in this architecture?
A) Scalability issues
B) Data consistency
C) Cost overhead
D) (your own answer)

The agent still gives its usual recommendation per grilling's rules, but the options are always lettered and free-form answers are welcome.

Compare before adopting

Before integrating grilling-fix, compare it against your current setup. Check quality signals like GitHub stars (currently 20) and recent commits. Evaluate alternatives—maybe you need a full grilling rewrite instead of a format fix. Also consider workflow fit: does your agent already handle free-text responses gracefully? If not, this skill alone won't solve parsing issues.

Why it is worth tracking

grilling-fix is a focused, low-risk enhancement that improves user experience without overhauling your agent's logic. Its niche scope means low maintenance, but its community momentum is still early. Track it if you use grilling and find that users struggle with numbered options or want more expressive answers. For more context, see the skill page for updates and community feedback. Evaluate it when you next refine your agent's interaction patterns—it might be the small polish that makes your grilling sessions feel more human.