w-grill — Claude Code skill: what it does, how to install
The w-grill Claude Code skill stress-tests a plan or decision by mapping its dependent choices. It asks about decisions whose prerequisites are settled, then waits for confirmation.
MIT 543 installs ★ 1 updated 2026-10-02
npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a claude-codeWhat w-grill does
w-grill examines a plan or decision by asking about the choices connected to it. First, it checks relevant MDX files, ADRs, the README, and DESIGN.md for facts already recorded. For filesystem and code facts, it follows the first three steps of w-council.
The skill maps decisions as a tree and finds the frontier: choices whose prerequisites have been settled. It asks about the whole frontier in one round, using references/round-format.md, then waits for answers and recalculates. Once no open questions remain, it confirms shared understanding. It does not carry out the plan until the user agrees. Automatic invocation is disabled.
What it helps with
- find unresolved choices in a plan
- group questions by decision dependencies
- hold implementation until the plan is confirmed
Example prompts for the agent
- Use w-grill to map the decisions in this plan and ask about the open frontier.
- Stress-test this proposal; check project files first, then ask the settled questions together.
How to install w-grill
1Skills CLI (Vercel)
npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a claude-code2Manually: put the files in place
~/.claude/skills/w-grill/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/w-grill/
1Skills CLI (Vercel)
npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a codex2Manually: put the files in place
.agents/skills/w-grill/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a cursor2Manually: put the files in place
.cursor/skills/w-grill/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install blockmatic-icebox/basilic-skills skills/workflow/w-grill2Manually: put the files in place
.github/skills/w-grill/SKILL.md1Skills CLI (Vercel)
npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a gemini-cli2Platform CLI
gemini skills install https://github.com/blockmatic-icebox/basilic-skills --path skills/workflow/w-grillManually: put the files in place
.devin/skills/w-grill/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about w-grill
How do I install w-grill in Claude Code?
Run npx skills add blockmatic-icebox/basilic-skills --skill w-grill -a claude-code. The skill file is at ~/.claude/skills/w-grill/SKILL.md.
What does w-grill do?
It maps the decisions behind a plan, asks about choices whose prerequisites are settled, and waits for confirmation before acting on the plan.
Which project files does w-grill inspect?
It checks relevant MDX files, ADRs, the README, and DESIGN.md. It uses the first three w-council steps for filesystem and code facts.
Source text
SKILL.md · MIT
--- name: w-grill description: Relentless design-tree interview until the frontier is empty. disable-model-invocation: true --- Stress-test a plan or decision. Blocking gaps only use [clarify](../w-clarify/SKILL.md). Do not implement until the user confirms shared understanding. Do not write a glossary or a second ADR tree. 1. Inspect matching MDX, ADRs, README, and `DESIGN.md` for facts already recorded. Follow [council](../w-council/SKILL.md) steps 1–3 for filesystem or code facts. Do not ask the user for anything you can look up. 2. Map the work as a **design tree**: every decision branches into the decisions that hang off it. The **frontier** is every decision whose prerequisites are already settled. 3. Ask the whole frontier in one round. Follow [round format](references/round-format.md). Then wait. 4. Each round of answers reshapes the tree. Recompute the frontier. A running exploration is an unsettled prerequisite: ask the rest of the frontier now; questions downstream of it wait. 5. Stop when the frontier is empty. Confirm shared understanding. Do not act on it until the user says to.
© blockmatic-icebox · MIT · blockmatic-icebox/basilic-skills