@santiagodevrel/arkiv-advisor-skill
v0.1.3
Published
Claude Code skill that helps builders make architecture and ideation decisions on Arkiv. Covers mental model, brainstorming walkthroughs, ETHLisbon sponsor mix patterns, business model patterns, and 30+ idea seeds. Pairs with arkiv-ethlisbon-skill for ful
Maintainers
Readme
Arkiv Advisor Skill
The WHAT/WHY layer of the Arkiv co-pilot stack — architecture brainstorming, idea generation, ETHLisbon sponsor combos, and business model patterns. Pairs with the official arkiv-best-practices (engineering) and arkiv-ethlisbon-skill (templates) for a full Claude Code co-pilot.
Install — one command, full stack
npx @santiagodevrel/arkiv-advisor-skill initThis installs 3 skills in one shot:
arkiv-best-practices(Arkiv official) — SDK reference, CRUD, 14 best practicesarkiv-ethlisbon-skill— 6 common pitfalls + 3 starter templates + demo prep checklistarkiv-advisor-skill(this) — architecture brainstorming + 30+ ideas + sponsor combos
Open a fresh Claude Code session — all three auto-load when Arkiv context is detected.
⚠️
npm idoes NOT install the skills. This package is a CLI. Usenpx ... initto install the skill files into.claude/skills/.
The 3-skill co-pilot stack
| Skill | What it covers | When Claude uses it |
|---|---|---|
| arkiv-best-practices (official) | SDK setup, CRUD, 14 best practices, mental model | When you write Arkiv code |
| arkiv-ethlisbon-skill | 6 common pitfalls, 3 starter templates, demo prep, prompting patterns | When you build / debug ETHLisbon-style apps |
| arkiv-advisor-skill (this) | Architecture brainstorming, 30+ idea seeds, sponsor combos, business models | When you decide WHAT to build / WHY |
Loaded together → Claude becomes a full Arkiv co-pilot: architecture, code, ideation, and strategy.
What this skill adds (the WHAT/WHY layer)
The advisor skill teaches Claude to help with strategic decisions, not just code.
Architecture brainstorming walkthrough
7 questions Claude walks through with you in plain language:
- Can the data change after I create it? (vote vs profile bio)
- How long does it need to live? (clipboard vs user profile)
- Who controls it? (
$ownercan transfer,$creatoris permanent) - Who pays for and signs writes? (user's MetaMask vs your backend wallet)
- More writes or more reads? (analytics dashboard vs profile registry)
- Are entities related? (parent-child via shared attributes)
- How big should my batch writes be? (50–200 entities per tx)
Plus 5 pre-armed architecture templates: read-heavy public app, write-heavy backend, MetaMask user-owned, event-sourced indexer, time-bounded coordination.
Pro-Arkiv competitive orientation
Brief paragraphs on Ceramic, Tableland, IPFS+OrbitDB, Lit Protocol, Arweave, ComposeDB. Never recommends leaving Arkiv — instead identifies when to combine (e.g., "Arkiv + Lit for encrypted queryable data", "Arkiv + IPFS for hybrid blob storage").
ETHLisbon sponsor combos
12 high-leverage Arkiv + sponsor patterns based on documented ETHLisbon 2023 and 2025 editions:
- Arkiv + Olas → agent memory layer for autonomous Mechs
- Arkiv + Kusama → cross-chain coordination indexer
- Arkiv + Gnosis Pay × Aave → private spend tracker
- Arkiv + Filecoin Foundation → hybrid video archive
- Arkiv + Lit Protocol → encrypted private journal with queryable tags
- Arkiv + WalletConnect → multi-device user data
- Arkiv + Wormhole → cross-chain message archive
- Arkiv + Rootstock → Bitcoin-anchored bounty board
- Arkiv + Fuel → fast L2 + Arkiv-indexed history
- Arkiv + iExec → privacy-preserving compute + queryable archive
- Arkiv + ENS → portable Linktree owned by user
- Arkiv + Gnosis Chain / Safe → DAO proposal threads
Each combo: use case + why powerful + 36-hour feasibility.
30+ concrete idea seeds
Categorized by Arkiv pattern:
- Agent memory & AI (5 ideas)
- User-owned content (5 ideas)
- Decentralized indexing (5 ideas)
- Time-bounded coordination (5 ideas)
- Verifiable activity logs (5 ideas)
- Content + ownership transfer (5 ideas)
- Sponsor combos (7 ideas)
Each idea: one-line description + Arkiv primitive used + 36h feasibility (yes/maybe).
Business model patterns
Analyzed from real Arkiv-built apps (WebDB, FileDB, ImageDB, DrawIODB, UmamiDB, CopyPal). The common pattern: freemium + TTL pricing. Free tier = short TTL guest mode. Paid = wallet-mode with extended retention.
Decision-support prompts
7 prompts that work well when this skill is loaded:
- "Walk me through the architecture for [app idea]"
- "Critique my idea for Arkiv-fit"
- "Generate 5 ideas combining Arkiv with [sponsor]"
- "What's the business model for [app]?"
- "Help me decide between idea A and idea B"
- "Make my idea more Arkiv-shaped"
- "Stack 3 sponsors for my hackathon idea"
CLI flags
npx @santiagodevrel/arkiv-advisor-skill init [flags]
--project Install into ./.claude/skills (project-local)
--skip-official Skip arkiv-best-practices install
--skip-ethlisbon Skip arkiv-ethlisbon-skill install
--minimal Equivalent to both --skip-* (advisor only)
--dry-run Print actions without writingManual install (each skill independently)
# 1. Official engineering skill
npx skills add https://github.com/Arkiv-Network/skills --skill arkiv-best-practices
# 2. ETHLisbon templates + pitfalls
npx @santiagodevrel/arkiv-ethlisbon-skill init --skip-official
# 3. This advisor skill
curl -L https://raw.githubusercontent.com/SantiagoDevRel/arkiv-advisor-skill/main/SKILL.md \
-o ~/.claude/skills/arkiv-advisor.mdVerify it loaded
After install + Claude Code restart, ask Claude:
"Walk me through the architecture for a Telegram tutor bot that remembers user history."
Expected: Claude walks through the 7-step framework (entity shape, TTL choice, ownership, wallet model, etc.) and recommends a template shape, citing the agent memory pattern.
Acknowledgments
Built on top of the arkiv-best-practices skill from the Arkiv team and complements arkiv-ethlisbon-skill.
Built for the ETHLisbon hackathon community — sponsors mix patterns reflect the documented 2023 and 2025 editions.
License
MIT — see LICENSE.
