New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Solana Kit Skill Packs

by solanabr in solanabr/ai-kit

Tells the agent which optional skill pack in the Solana AI kit fits the work at hand when it is not installed yet, and how to offer it to you for approval.

MITAuto-check: notesAgent Workflows

Install Solana Kit Skill Packs

skills CLI
$ npx skills add solanabr/ai-kit --skill skill-packs -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install solanabr/ai-kit skill-packs --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/solanabr/ai-kit.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/skill-packs .claude/skills/skill-packs && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
skill-packs
GitHub stars
109
Token cost
~3.3k tokens
SKILL.md length
1,806 words
Files
1
Skills in repo
3
Repo updated
First seen
Licence
MIT

At a glance

Tells the agent which optional skill pack in the Solana AI kit fits the work at hand when it is not installed yet, and how to offer it to you for approval.

  • Starting Solana work that a pinned extension covers but that may not be installed
  • SKILL.md covers What the kit can promise about…, Work to packs, Packs that need a key and Offering rules
  • Calls bash; needs CIRCLE_API_KEY and CIRCLE_ENTITY_SECRET
  • Being asked which optional skill packs the kit offers

What it does

Four packs ship installed in the kit: solana-dev, auditor-skill, colosseum and anthropic-skills. Everything else arrives on demand, so you are not expected to know the catalogue. The agent notices from the work itself that a pack would help, names it and lets you decide. Core packs are never the answer here, since they are already present.

Packs come in two kinds, both recorded in skill-registry.json. Pinned extensions carry a reviewed commit, install through .claude/bin/skills.sh or /add-skill, and are refreshed by /update. Add-ons have no tier, install by the command in their own registry entry, are not pinned or scanned, and run only after the safe-ai-skill check returns proceed: true. The agent reads an add-on's safety line aloud first, because some are stale and some write hooks or a statusline into settings.json.

A work-to-pack table maps jobs to packs: qedgen for a Lean 4 proof that a program invariant holds, light-protocol for ZK compression of state, and eth-to-sol for porting Solidity contracts. The kit also covers landing pages and UI design, motion, NFTs, DeFi SDKs, Unity games, mobile, RPC and indexing, pitch and go-to-market, and data visualization.

When your agent uses it

  • Starting Solana work that a pinned extension covers but that may not be installed
  • Being asked which optional skill packs the kit offers
  • Deciding whether to install an unpinned add-on after reading its safety line

Example prompts

  • “Which skill packs does this kit have for building an NFT mint?”
  • “I'm porting a Solidity contract to Solana. Is there a pack for that?”
  • “What extras can I install for Unity games on Solana?”

Requirements

  • The solanabr ai-kit, with .claude/bin/skills.sh and skill-registry.json

What it can do on your machine

Read from SKILL.md and the folder at commit f9024a7. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • bash

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • CIRCLE_API_KEY
    • CIRCLE_ENTITY_SECRET
    • GMGN_API_KEY
    • GMGN_PRIVATE_KEY
    • RESEND_API_KEY
    • ELEVENLABS_API_KEY
    • MISTRAL_API_KEY
    • ARISTOTLE_API_KEY
    • ALCHEMY_API_KEY

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Solana Kit Skill Packs loads about 3.3k tokens when it runs. Until then it costs about 130 tokens; SKILL.md has 1,806 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~130
When it runs · the whole SKILL.md, loaded when a task matches
~3.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:66
    |empty` line and never a value. Reading `.env` is denied at the Medium and High firewall tiers, and a value read into th

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from solanabr/ai-kit at commit f9024a7, republished under its MIT licence (© solanabr). 1,806 words, ~3,273 tokens.

Download SKILL.mdSave it as .claude/skills/skill-packs/SKILL.md (or your agent's skills folder).
name
skill-packs
description
Which optional skill pack the kit has for the work at hand, and how to offer it. Read it when starting work the kit pins a pack for and that pack may not be installed yet — program development and formal verification, landing pages and UI design, motion and animation, NFTs, DeFi SDKs, Unity games, mobile, RPC and indexing, pitch and go-to-market, data visualization — or when the user asks what packs exist. Covers the pinned extensions and also the unpinned add-ons that only skill-registry.json records.
user-invocable
true

Which skill pack for this work

Four packs ship installed — solana-dev, auditor-skill, colosseum and anthropic-skills — and everything else arrives on demand. The point of this file is that the user does not have to know the catalogue: notice from the work itself that a pack would help, name it, and let them decide. A core pack is never the answer here, because it is already there; this file only covers what still has to be fetched.

What the kit can promise about a pack

Pinned extensionAdd-on
In skill-registry.json"tier": "extension"no tier
Installbash .claude/bin/skills.sh add <id>, or /add-skill <id>the install.command in its registry entry — read the entry rather than reconstructing the command, because the method varies (submodule, clone, npx, a plugin marketplace)
Commit installedthe commit its registry entry records — reviewed here, and checked against the gitlink by validate.shwhatever upstream HEAD is that day, reviewed by nobody here
Kept by /updateyesno
Who runs itthe user, on a yesthe user, and only after safe-ai-skill add skill|mcp <source> returns proceed: true

An add-on entry's safety line is load-bearing, so read it out before the user installs: none of them is pinned or scanned at a commit, a few have been stale for a year, and some write hooks or a statusline into settings.json when run through their own installer.

Work to packs

Ids in the middle column install with bash .claude/bin/skills.sh add <id>. Ids in the right column need their registry entry.

WorkPinned extensionsAdd-ons
Any program work: Anchor, Pinocchio, nativeqedgen (Lean 4 proof that the invariant you just wrote holds)—
State that would cost too much as ordinary accountslight-protocol (ZK compression: compressed PDAs and compressed tokens, and the cost model that justifies them — last moved 2026-06, so check its SDK versions)—
Porting a Solidity or EVM contract—eth-to-sol (the Solana Foundation's type, pattern and stdlib mappings)
Landing page, marketing site, any UI surfacevercel (web design review, Next.js and React performance), get-shit-pretty (brand identity and a design system: shadcn theme, design tokens, UI critique, WCAG audit — 45 skills, 13 sub-agents), solana-new (brand design, design taste, frontend design guidelines, number formatting)anydesign (an image, URL or Figma frame into design tokens), design-skills (design critique, accessibility audit, journey mapping), ux-writing-skill (onboarding, error and empty-state copy)
Animation, motion, transitions, page-load choreographygsap-skills (GreenSock's own eight skills: timelines, ScrollTrigger, useGSAP, plugins, reduced-motion, performance — the one to offer when the ask is real motion work), solana-new (page-load animation and video-craft references)animation-principles (motion principles as prose; unmaintained since December 2025)
NFTs, collections, compressed NFTsmetaplex—
A token launch, airdrop, ToS or privacy policy, licensing or sanctions questioncrypto-legal (statutory citations across US, EU/MiCA and Brazil — informational only, never legal advice, and its review is pinned at 2026-06, so say both when you offer it)—
Swaps, lending, perps, oracles, bridgesjupiter, sendai, meteora-invent (Meteora's own skill — DBC, DAMM, DLMM, vaults, locks, M3M3 — deeper than the sendai folder, which it supersedes)—
Taking payment in USDC or a stablecoin, settling off a DEXcircle (18 skills: agent wallets, CCTP bridging, Gateway, Arc, accepting payments; needs CIRCLE_API_KEY and CIRCLE_ENTITY_SECRET). It moves real money — say so when you offer it, and point it at a testnet first—
Researching a memecoin or a wallet before touching itgmgn (17 skills: due diligence, holder and wallet analysis, smart-money tracking, dev score, narrative; needs GMGN_API_KEY, and GMGN_PRIVATE_KEY for orders). Two of its skills submit trades, and the swap one takes contract addresses only — never names — so flag both when you offer it—
Unity, C#, PSG1, real-time gamessolana-game, magicblock (ephemeral rollups for real-time state)—
React Native, Expo, Seeker, dApp Storesolana-mobile (MWA, Seeker, Genesis Token), expo (the other half: EAS Build/Update/Workflows CI, store submission, OTA, config plugins and native modules, SDK upgrades — offer both when the app is Expo)ios-simulator-skill, swiftui-design-skill
RPC, DAS, webhooks, indexing, edge hostinghelius, alchemy, cloudflare—
Browser QA of a dApp that is already runningplaywright-skill (a suite that has to live: persistent sessions, multiple contexts, CI — the core webapp-testing skill already covers a one-off check)dev-browser
An MCP server for your program or APIcloudflare (Agents SDK, MCP servers on Workers)anthropic-claude-code-plugins (its mcp-server-dev plugin: deployment models, tool design, auth)
Pitch deck, demo day, hackathon, competitive researchstartup-builder (pitch-deck and hackathon: the story, the track, the submission and the demo script), frontend-slides (the deck itself: 36 HTML templates, an intent→template selection index, PDF export and Vercel publish), solana-new—
Startup work around the project: validating an idea, build and traction status, positioning, pricing, fundraising, launch day, incident comms, listings and ecosystem BDstartup-builder (ten Markdown skills handing off through .claude/context/ files; idea-sprint writes the idea.md that /plan-feature reads). Legal questions it routes to crypto-legal—
Any slide deck, marketing graphic or social imagefrontend-slides (HTML is the house format for these; content-gen hands it the content)—
Getting a landing page, docs page or post found: SEO, keywords, search intent, rankingsuperseo (page audit, content brief, E-E-A-T scoring, topic clusters, link building — markdown only, and it uses your own search tools rather than a paid SEO API)—
Charts, dashboards, on-chain data visualization—claude-d3js-skill, scientific-agent-skills
Analysing or reporting on exported data: a Parquet or CSV dump, an airdrop snapshot, indexer outputduckdb (SQL over files with no database to stand up; supabase instead when the data has to live somewhere)—
A historical or cross-chain question someone has already decoded the data fordune (DuneSQL over decoded Solana and EVM tables; an indexer is still the answer for your own program's state)—
The datastore under an indexer or webhook consumersupabase (Postgres, Auth, Realtime), mongodb (documents, Atlas Search and vector search, stream processing)—
Caching, queues or rate-limit state on a hot pathredis (pooling and exhaustion, clustering, Redis Search, semantic caching for LLM calls)—
Standing up or changing the servers under the projectpulumi (infrastructure-as-code past where CI and Cloudflare stop; it drives the pulumi CLI against live cloud credentials), google (GKE, BigQuery, IAM, PromQL — about 20 skills run the gcloud, kubectl and bq CLIs against your active project)—
Email a dApp has to send: alerts, receipts, a waitlistresend (React Email, deliverability, an agent inbox; needs RESEND_API_KEY and sends as your domain, so a mistake reaches real inboxes)—
Narration, dubbing, sound or music for a demo or trailerelevenlabs (TTS, dubbing, sound effects, music, transcripts — every call is billed to the user's ELEVENLABS_API_KEY)—
The business around the project rather than the chain: sales, finance, support, HR, enterprise searchknowledge-work (252 skills across 18 role plugins; its 186 catalogued MCP connectors stay inert until the user adds one). For statutory questions use crypto-legal instead—
Research notes, a knowledge base, an Obsidian vaultobsidian-skills (Obsidian Markdown, Bases query views, JSON Canvas, the Obsidian CLI, Defuddle web-to-Markdown — offer it for a vault the user already keeps, not as a place to put project docs; vault-write and local-exec, see its registry safety field)—
An AI or LLM feature: a model, inference endpoint, fine-tune or Gradio demohuggingface-skills (the hf CLI and Hub, Gradio and Spaces, ZeroGPU, fine-tuning, evals, transformers.js, SageMaker — the one machinery pack, so name its registry safety field when offering it)—
Show full SKILL.md (619 more words)Show less

Some work needs no offer at all, because the pack for it already ships. A security audit, a dependency or secret scan, a threat model has no row above: read the core auditor-skill pack, which superseded the trailofbits, ghostsecurity, defending-code and safe-solana-builder packs the kit used to pin. General program, client and test work rests on core solana-dev — the program row above adds only the specialists on top of it. A visual direction for a UI, and Playwright QA of a running app, are core anthropic-skills: its frontend-design and webapp-testing install top-level and are listed every session, so they are already in front of you. Its third skill, mcp-builder, is not in the pack any more — Anthropic ships it per-user from its own marketplace. Colosseum archives, checking an idea against past submissions, and competitive research are core colosseum, which signs in through its own helper rather than an API key, so the key section below does not list it (the hub has the sign-in). Offering any of these is a wasted question.

The registry holds more entries than this table: watchlist items, archived repos, and duplicates of a pack already listed here. Search it by domain or tags when the work fits no row.

Packs that need a key

Name the key when offering the pack, and check whether it is set with bash .claude/bin/env-keys.sh, which prints one KEY set|empty line and never a value. Reading .env is denied at the Medium and High firewall tiers, and a value read into the transcript has already left the machine.

  • qedgen — MISTRAL_API_KEY for fill-sorry and generate, ARISTOTLE_API_KEY for the aristotle commands. Without either, its Lean references still read but nothing generates, so the pack is worth installing only alongside a key.
  • alchemy — ALCHEMY_API_KEY for the alchemy-api skill; its agentic-gateway skill is keyless (x402), so the pack still earns its place without a key
  • circle — CIRCLE_API_KEY and CIRCLE_ENTITY_SECRET. Several of its skills also read *_PRIVATE_KEY variables, and with those set it moves real money: name that, and point it at a testnet before anything else
  • gmgn — GMGN_API_KEY for every call, plus GMGN_PRIVATE_KEY (a PEM request-signing key for GMGN's API, not a chain wallet key) before its two order-submitting skills work. The research skills are useful with the first key alone, which is the safer thing to offer
  • elevenlabs — ELEVENLABS_API_KEY. Every generation call is billed to it, so a long dubbing or music job costs real money; say so rather than starting one
  • resend — RESEND_API_KEY. It sends as your own domain, so a mistake lands in real inboxes under your sending reputation.

A pack that is inert without a key the user has not set is a question, not an install: say which key it needs and what it would do once set.

Offering rules

  • One or two packs, with a sentence each on what they add that the kit does not already have. The user asked for the work, not for a shopping list.
  • Already installed (bash .claude/bin/skills.sh list marks it) means read it and carry on, with no offer. bash .claude/bin/skills.sh uninstalled is the list still worth offering.
  • Install on a yes, never ahead of one: it is a network fetch and a write under .claude/skills/. An extension goes through /add-skill <id>; an add-on's command is the user's to run.
  • A no stands for the rest of the session. Do the work with what is installed.
  • No .claude/bin/skills.sh means a plugin install or a kit older than the core/extension split, where no extension is installable. Route to the upstream marketplace or the full install instead, as the hub describes.
  • Which files to read once a pack is installed is the hub's job, not this file's. This one only answers which pack.

© solanabr, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/skill-packs of solanabr/ai-kit.

Open the folder on GitHubat commit f9024a7

Compare with similar skills

Solana Kit Skill Packs next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Solana Kit Skill Packs compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Solana Kit Skill Packs this skillsolanabr/ai-kit109—~3.3kAutomated safety check: NotesMIT
Darwin Skill Optimizeralchaincyf/darwin-skill6.2k1 repos~4.7kAutomated safety check: PassMIT
Using Agent Skillsaddyosmani/agent-skills104k4 repos~2.4kAutomated safety check: PassMIT
Ponytail Help CardDietrichGebert/ponytail160k—~726Automated safety check: PassMIT
Skill Creatorzhayujie/CowAgent47k—~4.7kAutomated safety check: NotesMIT
Task Observerrebelytics/one-skill-to-rule-them-all3.2k1 repos~11kAutomated safety check: PassCC-BY-4.0

Similar skills

  • Darwin Skill Optimizer

    alchaincyf/darwin-skill

    Scores SKILL.md files on a nine-dimension rubric, then improves them in a keep-or-revert loop with independent judge agents, test prompts, git history and human checkpoints.

    6.2k GitHub starsUsed in 1 repo~4.7k tokens
    Agent WorkflowsAuto-check passed

  • Using Agent Skills

    addyosmani/agent-skills

    Meta-skill for choosing which workflow skill fits the task at hand, plus always-on habits: surface assumptions, stop on confusion, push back, keep it simple and stay in scope.

    104k GitHub starsUsed in 4 repos~2.4k tokens
    Agent WorkflowsAuto-check passed

  • Ponytail Help Card

    DietrichGebert/ponytail

    Shows a one-shot quick-reference card for the ponytail skills: intensity levels, the six commands, and how to turn it off, set a default mode and update.

    160k GitHub stars~726 tokensUpdated today
    Agent WorkflowsAuto-check passed

  • Skill Creator

    zhayujie/CowAgent

    Guides creating, installing and updating agent skills in a workspace: SKILL.md frontmatter, bundled scripts and references, with scripts to scaffold, validate and package.

    47k GitHub stars~4.7k tokensUpdated today
    Agent WorkflowsAuto-check: notes

  • Task Observer

    rebelytics/one-skill-to-rule-them-all

    Monitors task execution for skill improvement opportunities.

    3.2k GitHub starsUsed in 1 repo~11k tokens
    Agent WorkflowsAuto-check passed

  • Open-Science Skill Creator

    aipoch/open-science

    Creates, revises, evaluates and publishes skills in the Open-Science app through its native host.skills composer, with optional test prompts and benchmarks.

    5.5k GitHub stars~1.7k tokensUpdated today
    Agent WorkflowsAuto-check passed

More from solanabr/ai-kit

  • Solana Token Extensions

    solanabr/ai-kit

    Helps choose, combine and create Token-2022 mint and account extensions on Solana with the spl-token CLI, @solana/kit or Anchor, and integrate the resulting mints.

    109 GitHub stars~3.5k tokensUpdated today
    Auto-check passed

  • Solana Dev

    solanabr/ai-kit

    Routing hub for Solana development. An agent skill from solanabr/ai-kit.

    109 GitHub stars~12k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Solana Kit Skill Packs

What does Solana Kit Skill Packs do?

Tells the agent which optional skill pack in the Solana AI kit fits the work at hand when it is not installed yet, and how to offer it to you for approval. Four packs ship installed in the kit: solana-dev, auditor-skill, colosseum and anthropic-skills. Everything else arrives on demand, so you are not expected to know the catalogue.

When should I use Solana Kit Skill Packs?

Solana Kit Skill Packs fits situations like: starting Solana work that a pinned extension covers but that may not be installed; being asked which optional skill packs the kit offers; deciding whether to install an unpinned add-on after reading its safety line.

How do I install Solana Kit Skill Packs in Claude Code?

Run `npx skills add solanabr/ai-kit --skill skill-packs -a claude-code`. Or copy the skill folder (.claude/skills/skill-packs in solanabr/ai-kit) into .claude/skills/skill-packs in your project. Claude Code loads it when a task matches its description.

How do I install Solana Kit Skill Packs in Codex?

Run `npx skills add solanabr/ai-kit --skill skill-packs -a codex`. Or copy the skill folder (.claude/skills/skill-packs in solanabr/ai-kit) into .agents/skills/skill-packs in your project. Codex loads it when a task matches its description.

Can I use Solana Kit Skill Packs in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add solanabr/ai-kit --skill skill-packs -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/skill-packs, .gemini/skills/skill-packs, .github/skills/skill-packs and .opencode/skills/skill-packs in your project.

What does Solana Kit Skill Packs need to run?

Going by SKILL.md and its folder, Solana Kit Skill Packs needs the command-line tools its instructions call (bash) and credentials named CIRCLE_API_KEY, CIRCLE_ENTITY_SECRET, GMGN_API_KEY and GMGN_PRIVATE_KEY. Our summary lists: The solanabr ai-kit, with .claude/bin/skills.sh and skill-registry.json.

Does Solana Kit Skill Packs access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Solana Kit Skill Packs safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Solana Kit Skill Packs use?

Solana Kit Skill Packs is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Solana Kit Skill Packs use?

About 3.3k tokens (SKILL.md is roughly 13k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Solana Kit Skill Packs?

Skills that share tags, products or a category with Solana Kit Skill Packs: Darwin Skill Optimizer (alchaincyf/darwin-skill, 6.2k stars), Using Agent Skills (addyosmani/agent-skills, 104k stars), Ponytail Help Card (DietrichGebert/ponytail, 160k stars) and Skill Creator (zhayujie/CowAgent, 47k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Solana Kit Skill Packs?

solanabr (a GitHub organization) maintains it in solanabr/ai-kit, which has 109 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on October 10, 2026.

Source: solanabr/ai-kit on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.