Skip to content
Conduit

skill

agent-skills

Agent Skills for Claude Code & skills.sh — senior engineering-manager skill for Next.js 16 + React 19 + Supabase + Vercel

darasoba/agent-skillscommunitylistedunreviewed

Set up

Skills are installed by placing their folder where your agent loads skills.

The download comes from github.com/darasoba/agent-skills itself, not from our server, so what you get is exactly what its author published.

Install instructions come from the README

Skills do not use an MCP config. Follow the steps in the README below.

README

synced Oct 26, 2026

agent-skills

Agent Skills for Claude Code and other skills.sh-compatible coding agents.

Skills

engineering-manager

A senior engineering-manager / principal-engineer persona for modern full-stack web apps. It owns architecture, holds the quality bar, and acts as the final reviewer before code ships — with deep, current (2026) command of the stack: Next.js 16 App Router, React 19, TypeScript, Tailwind CSS 4, Supabase (@supabase/ssr + RLS), Vitest, Playwright, and Vercel (Fluid Compute).

Use it to design a feature or system, make an architecture/tradeoff call, review or approve a change before merge, scope work, or get oriented in an unfamiliar codebase. It encodes the 2026 stack traps that stale knowledge gets wrong (opt-in caching, proxy.ts, Server Actions as public endpoints, RLS as the security boundary, Fluid Compute over Edge) and ships with a correctness-first review rubric.

fable-reasoning

Elite, stack-agnostic reasoning and coding discipline — a process layer that raises problem-solving quality on any task, in any language. It enforces hypothesis-driven thinking, evidence-before-belief, root-cause (not symptom) fixes, calibrated effort, and adversarial self-verification before anything is called "done."

Load it at the start of any non-trivial task — debugging, implementing a feature, refactoring, an architecture decision, code review, or a multi-step investigation. The main SKILL.md carries a 6-phase operating loop (understand → gather context → hypothesize → plan → execute → self-verify), an effort-calibration table, and an anti-pattern list; three reference files go deep on debugging, coding craft, and verification, loaded only when the task centers on that activity.

Install

With the skills CLI (no install needed):

# Add every skill in this repo to the current project
npx skills add darasoba/agent-skills

# Or install globally (user-level), available in every project
npx skills add darasoba/agent-skills --global

# List what's in this repo without installing
npx skills add darasoba/agent-skills --list

This works with Claude Code and any other agent the skills CLI supports.

Layout

skills/
├── engineering-manager/
│   ├── SKILL.md                       # persona, operating principles, stack traps, routing
│   └── references/
│       ├── nextjs-react.md            # Next 16 / React 19: boundaries, Server Actions, caching, perf
│       ├── supabase-vercel.md         # @supabase/ssr, RLS, key migration, Fluid Compute, env/cron
│       ├── testing-tailwind-zod.md    # Vitest pyramid, tz-deterministic testing, Tailwind 4, Zod 4
│       ├── codebase-navigation.md     # getting oriented in a Next.js + Supabase repo
│       └── review-rubric.md           # the ship-gate checklist
└── fable-reasoning/
    ├── SKILL.md                       # 6-phase reasoning loop, effort calibration, anti-patterns
    └── references/
        ├── debugging.md               # reproduction, bisection, hypothesis trees, heisenbugs
        ├── coding.md                  # reading order, idiom-matching, edge-case taxonomy, safe refactors
        └── verification.md            # evidence hierarchy, discriminating tests, adversarial self-review

License

MIT © Dara Sobaloju

↑↓ navigate openesc close