Interrogate plans and designs to expose weak assumptions and unclear decisions.
npx skills add https://github.com/dmmulroy/skills --skill grill-me-dmmulroy
Please help me install this Agent Skill. Skill: grill-me Source: https://github.com/dmmulroy/skills/tree/main/grill-me Command: npx skills add https://github.com/dmmulroy/skills --skill grill-me-dmmulroy
This Skill helps you pressure-test a plan or design before you commit to it, exposing weak assumptions, missing details, and unclear decisions early.
Use the grill-me skill to challenge my current plan until the assumptions, risks, and next steps are clear.
High-intent search queries and answers about installing and using this skill.
To pressure-test a technical design, you interrogate the plan through a relentless question loop until reasoning is specific, exposing weak assumptions, missing details, and unclear decisions before execution.
Finding hidden risks in a draft architecture involves applying sharp follow-up questions to surface gaps in scope, sequencing, dependencies, and tradeoffs, ensuring the roadmap is defensible before team presentation.
Yes, you can use relentless questioning to clarify scope and dependencies for a software refactor plan by pushing the proposal with evidence-backed clarification until all risks and next actions are explicitly resolved.
Validating implementation proposals through a design critique interview requires demanding evidence-backed clarification for each decision, pushing the proposal until vague reasoning becomes specific and defensible.
You need a questioning loop to review an architecture plan when you have a draft design requiring pressure-testing to tighten it up, ensuring hidden risks are found and assumptions are clear before committing to execution.
psengerIdentify design risks in architectures and plans using pre-mortem, red teaming, and ATAM techniques.
shaman007Identify and resolve ambiguities in feature specifications through structured questioning.
dthuy62Create detailed technical implementation plans with phases and dependency tracking.
Chang-Shih-YungAggregate type checker, linter, and test results into a 0-10 composite score.
