By ahtisamafzal
Karpathy behavioral principles + Pocock engineering workflow skills
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.
Set up git safety guardrails to block dangerous git commands (push, reset --hard, clean, branch -D, etc.) before they execute. Supports Claude Code and opencode. Use when user wants to prevent destructive git operations or add git safety hooks.
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.
Build a throwaway prototype to flesh out a design before committing to it. Routes between two branches — a runnable terminal app for state/business-logic questions, or several radically different UI variations toggleable from one route. Use when the user wants to prototype, sanity-check a data model or state machine, mock up a UI, explore design options, or says "prototype this", "let me play with it", "try a few designs".
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Karpathy behavioral principles + Pocock engineering workflow skills, packaged as a single installable skill set for Claude Code and opencode.
These are active in every conversation without any command:
| Skill | What It Does |
|---|---|
/diagnose | Disciplined diagnosis loop for hard bugs and performance regressions |
/git-guardrails | Block dangerous git commands (push, reset --hard, clean, branch -D) |
/grill-with-docs | Grilling session that challenges your plan against domain model and docs |
/improve-codebase-architecture | Find deepening opportunities informed by domain language and ADRs |
/prototype | Build a throwaway prototype to flesh out a design |
/review | Review changes against standards and spec side by side from a fixed point |
/setup-combined-skills | Scaffold per-repo config (issue tracker, triage labels, domain docs) |
/setup-pre-commit | Set up Husky + lint-staged + Prettier pre-commit hooks |
/tdd | Test-driven development with red-green-refactor loop |
/to-issues | Break a plan into independently-grabbable issues using vertical slices |
/to-prd | Turn conversation context into a PRD and publish to issue tracker |
/triage | Triage issues through a state machine driven by triage roles |
/zoom-out | Get broader context or a higher-level perspective on code |
| Skill | What It Does |
|---|---|
/caveman | Ultra-compressed communication mode (~75% fewer tokens) |
/grill-me | Relentless interview about a plan or design |
/handoff | Compact conversation into a handoff document for another agent |
/write-a-skill | Create new agent skills with proper structure |
| Karpathy Principle | Pocock Skill | How They Reinforce Each Other |
|---|---|---|
| Think Before Coding | grill-me, grill-with-docs | When uncertain, interview rather than assume |
| Simplicity First | to-issues, tdd | Vertical slicing keeps each slice minimal |
| Surgical Changes | zoom-out, improve-codebase-architecture | Understand context first; deepen modules surgically |
| Goal-Driven Execution | tdd, diagnose, setup-pre-commit | Automated verification loops on every commit |
| Platform | Always-on file | Skill discovery |
|---|---|---|
| Claude Code | CLAUDE.md | SKILL.md + plugin.json |
| opencode | AGENTS.md | SKILL.md in ~/.agents/skills/ |
After installation, run /setup-combined-skills in your project to configure the issue tracker, triage labels, and domain doc layout. This only needs to be done once per project.
npx skills@latest add ./path/to/combined-skills
Copy skills to ~/.agents/skills/combined-engineering-skills/ (opencode auto-loads from this directory):
cp -r skills/ ~/.agents/skills/combined-engineering-skills/skills/
Copy AGENTS.md to your project root or reference it in opencode.json:
{
"$schema": "https://opencode.ai/config.json",
"instructions": ["AGENTS.md"]
}
Restart opencode for changes to take effect.
/grill-me — it should activate and start interviewing you/setup-combined-skills — it should walk through issue tracker configurationRemove the skill set from your Claude Code configuration:
npx skills@latest remove combined-engineering-skills
Remove the skills directory and clean up AGENTS.md reference:
rm -rf ~/.agents/skills/combined-engineering-skills/
Then remove the instructions entry from your opencode.json or delete AGENTS.md from your project root.
| Category | Items | Reason |
|---|---|---|
| Personal skills | edit-article, obsidian-vault | Matt's personal setup |
| Misc skill | scaffold-exercises | Matt's specific course tooling (pnpm ai-hero-cli) |
| Deprecated skills | design-an-interface, qa, request-refactor-plan, ubiquitous-language | No longer maintained |
| In-progress skills | writing-beats, writing-fragments, writing-shape | Not yet ready |
| Personal tooling | migrate-to-shoehorn | Matt's personal tooling |
npx claudepluginhub ahtisamafzal/combined-engineering-skillsA growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC plugin for engineering teams - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
UI/UX design intelligence. 84 styles, 161 palettes, 73 font pairings, 25 charts, 17 stacks (React, Next.js, Vue, Nuxt.js, Nuxt UI, Svelte, Astro, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, Jetpack Compose, Angular, Laravel, JavaFX, Three.js). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.