# Ask Matt

> Pick the GrokBuild skill or flow that fits. Use when the user asks which skill to use, which workflow, "alur apa", "ask matt", or is choosing between planning, implement, review, and design paths.

- Skill: `kuker24/ask-matt` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add kuker24/ask-matt`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kuker24/ask-matt/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: kuker24 (https://skillmd.com/u/kuker24)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kuker24/ask-matt

---


<!-- opencode-bestfriend-overlay:ask-matt -->

# Ask Matt

Follow the thin router in `~/.config/opencode/AGENTS.md`. If routing is non-obvious, Read `~/.config/opencode/bestfriend/rules/00-routing.md`. Route only to skills installed here. Do not load every specialist. One primary specialist, plus at most one verification specialist when a risk or measured-performance trigger is on.

## Installed

**Plan:** `ask-matt` `grill-with-docs` `grilling` `to-spec` `to-tickets` `tdd` · architecture DAG = OpenCode plan agent

**Write:** this session. Matt ticket loop only: `/matt-implement`. Test-first: `tdd`.

**Review:** in-session (default). `matt-code-review` if they asked two-axis Standards + Spec. `/interrogate` for adversarial multi-review (manual).

**Design:** `found-this-design` `impeccable` `scroll-craft` `visual-studio` `scroll-world` `emil-design-eng`

**Documents:** `smartdoc` (per-job). `smartbook-ingest` only to compile/update reusable knowledge.

**Engineering (auto when the description matches):** `diagnosing-bugs` `domain-modeling` `codebase-design` `writing-for-agents` `research` `prototype`

**Engineering (manual / slash only):** `/architect` `/arena` `/blast-radius` `/why` `/figure-it-out` `/decision-log` `/improve-codebase-architecture` `/create-verification-skill` `/maintain-verification-skill` `/technical-writing` `/unslop` `/reflect` `/wizard` `/wait-what`

**Browser / GitHub / risk:** `browser-act` `chrome-devtools-axi` `gh-axi` `full-audit-keamanan` `full-performance-audit` `adhd`

**Not installed — say so, then use the nearest installed skill:** `/design` `/execute-plan` `/implement` `/review` `/code-review` `/imagine` `/docx` `/pdf` `/pptx` `/grill-me` `/handoff` `/triage` `/wayfinder` `/bro` `/poteto-mode` `/swarm` `/setup-matt-pocock-skills` `/pr-babysit` `/create-skill` `/create-workflow` `/build-with-ai` `game-asset-*`

## Route

1. Repo evidence is enough → do the work.
2. User typed a slash command → load it.
3. Architecture / PR-plan DAG → OpenCode plan agent, then implement in-session after approval.
4. Feature needs an interview, glossary, or ADR → `grill-with-docs`. Then `to-spec` → `to-tickets` only if they asked for tickets or the work is multi-session.
5. Ordinary implementation → write here. `tdd` when test-first. Do not start `/matt-implement` unless they are on a `/to-tickets` ticket.
6. UI world unknown → `found-this-design` then `impeccable`. World already chosen → `impeccable`.
7. Official library/docs fact → `research` (Context7). Why *this repo* chose an approach → `/why` (manual).
8. A missing Matt/Pstack name is the only named fit → say it is not installed. Use `grill-with-docs`, in-session write, in-session review, `adhd`, or `impeccable`.

## Phase boundaries

Continue if this session still holds the why. Start a new session if the window is disposable. Task tool for a scoped AFK task. OpenCode native compaction last.