bOpen.ai

Skills, Agents & Tools for Builders.

[bOpen.ai @bOpen_AI]

Want to learn about how to use skills, agents and tools for AI?

You should set alerts for this account!

We will be introducing you to our tools and characters regularly.

[BSV Association @BSVAssociation]

@bOpen_AI's Intro to BSV coding on Claude
► https://t.co/t3FxHHY1kK

[bOpen.ai @bOpen_AI]

Meet Wags, manager agent based on @danwag06.

Drop a feature description in Claude Code.

He creates the Linear issues, writes the user stories, structures the sprint, maps dependencies, and links it back to the roadmap.

No Jira ceremony. No ticket grinding. Just structure.

[bOpen.ai @bOpen_AI]

We're going to do an agent introduction sprint, friends!

Here's a few great agents that you should be using in your development stack and your business...

RIGHT NOW!

Paul closes security gaps with intent far better than Claude, Codex or Cursor.

🧵👇

[bOpen.ai @bOpen_AI]

Meet Ridd, our designer agent.

Tell him what component you need, and he configures shadcn, writes the Tailwind, handles dark mode, generates the SVGs and icons via Gemini Pro, and opens a https://t.co/dtoJvplMaQ mockup.

Design system to production component. One prompt.

[Cʀᴜᴍʙs @shadilayvision]

@BitcoinSVtrain @bOpen_AI

[bOpen.ai @bOpen_AI]

Caught in the Coinbase layoffs?

Our DMs are open.

bOpen is hiring. We build native L1 payments, identity, tokens, and applications that scale. For agents and humans.

If that's the work you want to do, send a message.

https://t.co/nJCx6tqhPv

[bOpen.ai @bOpen_AI]

One prompt covering Vercel, Postgres, OAuth, and rate limits at once will lose the thread.

Dispatch the right specialist with one tool call. Cleaner context. Sharper output. Parallel work when tasks are independent.

Named, illustrated, open source.

https://t.co/nJCx6tqhPv

[bOpen.ai @bOpen_AI]

What is a subagent?

In Claude Code, a subagent is a specialist you call inline. It runs in its own context & reads and tool calls never bleed back to the parent.

https://t.co/Hfm07megLB ships 39. Theo for Next.js. Idris for databases. Paul for sec ops. Maxim for integrations.

[bOpen.ai @bOpen_AI]

Most devs ask Claude to "write a Next.js app" and get training-data soup: React 18 patterns in a 19 codebase, Pages Router mixed with App Router, Client Components where Server Components run faster.

Theo has written rules. The same rules. Every time.

https://t.co/nJCx6tqhPv

[bOpen.ai @bOpen_AI]

Meet Theo.

Theo is https://t.co/Hfm07megLB's Next.js specialist. He carries explicit Skill files like Vercel best practices, React 19 async Server Components, Turbopack config, Bun and Biome setup, so you get consistent, opinionated output, not a different answer every session.

[bOpen.ai @bOpen_AI]

We shipped agent-readiness this week.

Any API or enterprise service can now query a https://t.co/Hfm07megLB agent's identity and capability state before granting access.

Well-known endpoints. Machine-readable.

No custom handshake.

→ https://t.co/nJCx6tqhPv

[bOpen.ai @bOpen_AI]

Most AI tools help you write code.
Jason helps you trust it.

Runs the AAA pattern on every test. Detects coverage gaps. Writes the edge cases and auth failures you forgot.

"We have tests" is not the same as "we have coverage."

→ https://t.co/lUqNi6tEaD

[bOpen.ai @bOpen_AI]

Meet Jason, the tester agent!

One prompt. He writes unit tests, integration tests, and Playwright e2e tests. Configures coverage reporting and wires it into CI/CD. Uses the Page Object Model. Handles mocking decisions.

Runs your repo. Against your actual code.

[bOpen.ai @bOpen_AI]

Wags covers sprint planning, PRDs, OKR sessions, roadmaps, retros, stakeholder maps, and pre-mortems.

All from a single prompt inside Claude Code. He's not a dashboard. He's the process layer for your whole team.

https://t.co/lUqNi6tEaD

[bOpen.ai @bOpen_AI]

Meet Wags, the bOpen project manager agent based on @danwag06

Drop a feature description in Claude Code.

He creates the Linear issues, writes the user stories, structures the sprint, maps dependencies, and links it back to the roadmap.

No Jira ceremony. No ticket grinding...

[bOpen.ai @bOpen_AI]

10 hook types. All defined in settings.json.

What we've shipped with them:
→ Lint check before every bash run
→ Auto-notify on file writes
→ Commit after session ends
→ Log every tool call to Bitcoin for audit.

The automation layer most people skip.

https://t.co/Hfm07megLB

[bOpen.ai @bOpen_AI]

Claude Code Hooks are the most underused feature in the agentic stack.

Every time Claude touches a file, runs a command, or sends a message, you can fire a shell script.

Before the action. After the action. Both.

No prompting. No babysitting. Just clean automation.