skillZs
★ LIVE SKILL TAGS ★
>>> LIVE SKILLS INDEX <<<
* OPEN SOURCE *
NO LOGIN, NO TRACKING
※ REAL INSTALL DATA ※
← back to all skills
wilbeibi/wilbeibi-skills73 installs

grill-me

Interview the user one question at a time until a plan is resolved, capturing durable decisions in CONTEXT.md or ADRs when the repo warrants it. Use when asked to "grill me," stress-test a design, or challenge assumptions before implementation.

How do I install this agent skill?

npx skills add https://github.com/wilbeibi/wilbeibi-skills --skill grill-me
view source ↗

Is this agent skill safe to install?

  • Gen Agent Trust Hubpass

    The skill interviews users to stress-test designs and document decisions. It is generally safe but has a low-risk attack surface for indirect prompt injection because it reads codebase files and documentation which could contain malicious instructions designed to influence the agent's behavior.

  • Socketpass

    No alerts

  • Snykpass

    Risk: LOW · No issues

What does this agent skill do?

grill-me

Relentlessly interview the user about a plan or design until you both reach shared understanding and every open decision is resolved.

How to run it

  • Read the plan, linked docs, code, or local files needed to understand the topic first.
  • Ask one question at a time. Wait for the answer before the next question.
  • For each question, state your recommended answer and a one-line why.
  • Walk each branch of the decision tree. When an answer opens new questions, follow them; resolve dependencies between decisions in order rather than jumping around.
  • If a question can be answered by exploring the codebase, explore instead of asking.
  • Push on fuzzy words, hidden assumptions, edge cases, dependencies between decisions, and reversibility.
  • Keep going until no unresolved branch remains. Then summarize the agreed plan.

Question patterns

  • "What outcome would make this change a success?"
  • "Which existing concept owns this behavior?"
  • "What is the smallest scenario that proves the design works?"
  • "What breaks if this assumption is false?"
  • "What would a future maintainer be tempted to change back?"
  • "Which term should be canonical, and which aliases should be avoided?"

Capturing decisions

Only when the repo already keeps CONTEXT.md, CONTEXT-MAP.md, or docs/adr/, or the interview resolves terminology or a decision worth outliving the conversation. Check for those files before proposing new ones.

  • Update the appropriate CONTEXT.md as soon as terminology is resolved — it is a glossary and relationship map for project-specific language, not a spec or scratchpad.
  • Use CONTEXT-MAP.md only when the repository has multiple bounded contexts.
  • Offer an ADR only for decisions that are hard to reverse, surprising without context, and based on a real tradeoff.
  • Create docs lazily: no CONTEXT.md or docs/adr/ until there is real resolved content, and keep them short enough that future agents actually read them.

For glossary format, read CONTEXT-FORMAT.md. For ADR format and thresholds, read ADR-FORMAT.md.

Notes

  • Do not soften or batch questions to "be efficient" — the value is the relentlessness. One question, one recommendation, repeat.
  • Stop early only if the user explicitly says to stop or accepts all remaining recommendations at once.
  • Ported from mattpocock/skills (MIT), adapted to this repo's conventions and voice.

Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.

<a href="https://skillzs.dev/skills/wilbeibi/wilbeibi-skills/grill-me">View grill-me on skillZs</a>