Grill
Grill a decision, plan or forecast. An outside AI judge, a model from a different company than Claude, argues the strongest case for and against it, names the cheapest test that would settle each challenge, and gives a verdict. Use when the user says grill, stress-test, pressure-test, poke holes in, red-team or challenge something they decided or plan to decide, when the weekly review marks a decision for grilling, or when they say look back or paste a decision record to score.
Install
npx skills add mtangoz/grill --skill grillgit clone --depth 1 https://github.com/mtangoz/grill.git /tmp/grill
mkdir -p ~/.claude/skills
cp -r /tmp/grill/skills/grill ~/.claude/skills/grillPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Grill a decision
The user has been thinking with you, so your critique is correlated with theirs. Grilling sends the decision to a judge from a **different model family**. Grill is the outside judge step Felix Rieseberg recommends in [How to be more ambitious](https://felixrieseberg.com/how-to-be-more-ambitious-a-tutorial-for-humans-with-models/): do not ask the model to judge itself. The judge writes:
- a steelman of the decision, and a counter-steelman;
- challenges, each with a **falsifier**: the cheapest test that would settle it;
- a verdict, in the same words as the website: solid, solid if, shaky, or doesn't hold up. The tool's own result uses `holds`, `holds-with-conditions`, `weak` or `refuted` for those four.
## First time
If the user seems new ("what is this?", "set up Grill"), say in two lines what it does. Then offer to grill **one real decision right now**: something they decided or will decide this week, and how sure they are. The first grill is the setup.
If they set up the one-click route with their own model-router key (not Grill Pro, whose account is already set this way), mention once the two settings in that account that keep write-ups unstored. Both are off by default:
- **Input & Output Logging** (Observability settings): leave it off, or list their Grill key under "Excluded API Keys".
- **Use of inputs/outputs** (Privacy settings, the 1% discount): leave it off.
## 1. Send the decision as statedDetails
- Skill name
grill- Repository
- mtangoz/grill
- Path
skills/grill/SKILL.md- License
- MIT
- Repo updated
- 30 October 2026
- Listed in
- karanb192/awesome-claude-skills
FAQ
What does the Grill skill do?
Grill a decision, plan or forecast. An outside AI judge, a model from a different company than Claude, argues the strongest case for and against it, names the cheapest test that would settle each challenge, and gives a verdict. Use when the user says grill, stress-test, pressure-test, poke holes in, red-team or challenge something they decided or plan to decide, when the weekly review marks a decision for grilling, or when they say look back or paste a decision record to score.
How do I install the Grill skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add mtangoz/grill --skill grill. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Grill popular?
It lives in mtangoz/grill, which has 1 GitHub stars and 0 forks, and is included in 1 curated skill list.
Related skills
Keyboard Shortcuts
nparashar150/claude-keyboard-shortcuts
Audit a web app for missing keyboard shortcuts and implement them end to end: a command registry, context-scoped bindings, a command palette, visible kbd hints, a ? help sheet, invariant tests, and WCAG 2.1.4 compliance. Use whenever the user mentions keyboard shortcuts, hotkeys, keybindings, a command palette, Cmd+K, power-user features, or making an app feel fast, and also when building or reviewing any list, table, modal, form, or editor UI where shortcuts and keyboard access obviously belong even if the user never says the word shortcut.
Never Again
1malaysherasia-ai/claude-never-again
Capture a bug or mistake as a permanent guardrail so it cannot happen twice. Use this skill whenever a bug is fixed, a test fails and is repaired, the user says "you did that again", "don't do that", "remember this", "never do X", or any correction is made that a future session would otherwise forget. Also use when the user asks to review, promote, or clean up existing lessons and hooks.
Prove It Ran
agentkitco/agentkit-dev-skills
Verify that tests, builds and checks actually ran and exercised the changed code, before claiming a task is done. Use before reporting completion, or when asked to prove tests ran or verify a fix.
River Trash Boom Robot Design
5dungnotnull/river-trash-boom-robot-design-agent-skill
Autonomous River Trash-Boom Design & Testing — Production-grade harness for Autonomous River Debris-Boom Robotics & Hydraulics evidence-backed analysis with real-time data aggregation, recognized domain methods, and self-improving knowledge pipeline.
Test Checklist
Ifeanyiejindu/qarunbook
Build a complete QA test plan for an app — every module, feature and user journey, across each platform it runs on. Use when asked for a test plan, QA checklist, test cases, regression suite, or "what should we test", or when preparing an app for QA, a release or a demo. Writes the plan, puts it into qarunbook (creating the app, or adding to it), then offers to start testing it straight away.
Lead Free Heat Resistant Ceramics
5dungnotnull/lead-free-heat-resistant-ceramics-agent-skill
Lead-Free Heat-Resistant Ceramic Glaze & Body Formulation — Production-grade Ceramic Materials & Lead-Free Glaze Engineering analysis harness. Use this skill whenever analyzing ceramic materials, formulating lead-free glazes, designing clay bodies, evaluating thermal expansion fit, testing food-safety compliance, or recommending firing schedules. This skill provides evidence-backed, risk-disclosed analysis with authoritative sourcing.
Browse the most popular Claude skills, all testing & qa skills, or write your own with the skills generator.