Why Skills v1.2 Is a Must-Have for AI Coding (Matt’s Top 5 Skills Lead the Leaderboard)
The new Skills v1.2 release adds Claude Code plugin support, revamps the grilling workflow to cut interaction rounds, and introduces three practical new skills—/wizard, /to-questionnaire, and /wait-what—while refactoring existing ones, offering a focused solution to the maintainability problems of AI‑generated code.
Four Core Pain Points Addressed by Skills
/grill-me – Clarify every ambiguous requirement before the agent proceeds.
grill-with-docs – Replace verbose agent output with concise documentation.
/tdd – Adopt a red‑green‑refactor loop to keep code runnable.
/improve-codebase-architecture – Continuously scan and deepen architectural improvements.
Claude Code Plugin Support
Version 1.2 adds official Claude Code marketplace installation.
claude plugins install mattpocock-skills /plugin install mattpocock-skillsGrilling Overhaul – From 13 Rounds to ~3
The original /grill-me over‑decomposed simple requests into dozens of sub‑questions. v1.2 replaces this with a single‑round questioning model: the agent asks all possible questions in one pass, recalculates the frontier, and proceeds to the next round. The result is roughly three rounds instead of thirteen.
Each round follows a fixed format:
❓ Q1 - <question title>
<question body or options>
➡️ <recommended choice>Answers can be given by number without re‑typing the full question.
Three New Skills
/wizard – Graphical Maintenance of Sensitive Information
Generates an interactive Bash script that guides the user through environment‑variable setup, API‑key configuration, and other sensitive operations, writing the results to .env and GitHub Actions secrets. The script runs model‑invoked, so the agent triggers it automatically when needed.
/to-questionnaire – Questionnaire Mode
Extracts essential decision‑making questions from the context and presents them ready to forward to a leader, eliminating the manual brainstorming step.
/wait-what – Trim Agent Verbosity
Three‑line skill that forces the agent to restate the problem concisely using ASD‑STE100 simplified technical English, leveraging shared vocabulary from CONTEXT.md.
Refactored Skills
prototype – Archiving of Prototype Diagrams
Preserves a single‑file HTML/CSS/JS prototype with a state panel and interactive walkthrough. After validation, the prototype is committed to a prototype/<name> branch, keeping the evidence while the main branch retains only decisions.
wayfinder – Multi‑Session Enhanced /grill-me
Scans the codebase, iteratively asks for goals and constraints, and turns unknowns into decision tickets. Sub‑agents perform automatic research; the user resolves pending tickets. The workflow is:
/wayfinder → /to-spec → /to-tickets → /implementSuitable for refactoring legacy systems or cross‑module features; not intended for small bug fixes.
writing-for-agents (renamed from writing-great-skills )
Extends documentation coverage to all files agents read (Skills, CLAUDE.md, AGENTS.md, etc.). Merges GLOSSARY.md into SKILL.md and moves technical details to SKILL-MECHANICS.md. The skill is model‑invoked during skill creation or modification.
Summary
The Skills suite directly addresses real‑world engineering pain points: misaligned requirements, non‑runnable code, and deteriorating architecture. It enables developers to work in their own style rather than being forced into a tool‑specific workflow.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Java Architecture Diary
Committed to sharing original, high‑quality technical articles; no fluff or promotional content.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
