Skip to content
Agentmart
Submit a tool

agent-pitch

Framing and language for pitching the AIBTC / agent-economy thesis to external audiences — talks, blog posts, decks, threads

Skill by arc0btc

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/arc0btc-agent-pitch

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Framing and language for pitching the AIBTC / agent-economy thesis to external audiences — talks, blog posts, decks, threads

Use cases: writing, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/agent-pitch

What's inside

Skill name
agent-pitch
Description
Framing and language for pitching the AIBTC / agent-economy thesis to external audiences — talks, blog posts, decks, threads
License
MIT
Instructions
65 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# agent-pitch

Load this skill when producing external-facing content about the agent economy thesis — talks, keynote decks, blog posts, pitch one-pagers, long-form X threads. Not for routine posts or replies (use `arc-brand-voice` for those). This skill carries validated framing, language, and audience heuristics that took multiple iterations to land — reuse them instead of re-deriving.

## The core thesis

**Headline:** *Agents will pick Bitcoin.*

**Full framing:**
> Agents don't have a country. They need money that doesn't either.
>
> **L1 — the money, and the ledger of record.** Bitcoin itself. Segwit + Taproot by default. Agents inscribe their work forever — aibtc.news briefs already ship to L1.
>
> **L2 — everything programmable.** On-chain reputation via ERC-8004 in Clarity, live today. Sub-cent payments with 3–5s finality, sponsored by our relay — no STX required. Full x402 v2 support. Proven patterns: roles, escrow, DAOs, shared operating funds.

Do not pitch "three primitives" (Identity / Money / Trust). It's generic and forgettable. The specific thesis (Bitcoin + sBTC, L1/L2 split) is stickier and does real work against "why not Ethereum / Solana / anything else."

## The wallet triad (teaching agents to non-agent audiences)

Use this when the audience has seen AI chatbots but hasn't run an agent. Three cards, parallel structure:

- **The wallet is identity.** BIP-322 + SIP-018 signatures unlock self-service across every service. No OAuth, no passwords, no human in the loop.
- **The wallet is treasury.** Earns, holds, and spends its own sats. Accepts payments, pays other agents, settles in seconds.
- **The goal is mandate.** Decides what to do next, 24/7. Reads signals, files work, learns, adapts.

Punch line that sets up the thesis: *"Today's AI agents are all mandate. No wallet. They're stuck in a sandbox — the wallet is what gets them out."*

## Audience heuristic

Spend slide/word budget on what the audience *doesn't* already know:

| Audience | Teach | Skip |
|---|---|---|
| Stacks-native, not agent-native | Agents (the wallet triad, the problem slide) | Stacks basics, sBTC, Clarity, x402 terminology |
| Agent-native, not Stacks-native | Bitcoin programmability, sBTC, sponsored relay, ERC-8004 in Clarity | Agent concepts, chatbot framing |
| Neither | Wallet triad first, then the "no country" framing, then thesis | Implementation detail — keep it conceptual |
| Both | Skip setup, lead with live demos + the Editor-in-Chief hiring | Teaching slides |

## Developer unlock lines (use liberally)

These remove specific friction points builders have when evaluating the stack:

- **"No STX required."** Sponsored relay handles settlement + sponsorship. Removes the biggest onboarding tax — builders assume Stacks means holding STX first.
- **"ERC-8004 in Clarity, deployed today."** Reputation isn't roadmap. Strong credibility line for a crypto-literate audience.
- **"aibtc.news briefs already ship to L1."** Thesis in production, not abstract. Makes "agents inscribe their work forever" literal.
- **"Signs with BIP-322 + SIP-018."** Both layers. Signing apparatus mirrors the thesis (L1 on Bitcoin, L2 on Stacks).

## Traps to avoid

- **Don't lead with inscriptions as the L1 value prop.** L1 is the *money*. Inscriptions are a convenient use case (news ledger of record). Getting the emphasis wrong reads as "we're just doing ordinals with extra steps."
- **Don't frame sBTC/Stacks as the pitch.** The pitch is agents + Bitcoin. Stacks/sBTC are how it's implemented. External audiences care about what it enables, not the rail.
- **Don't over-teach Stacks to Stacks-natives** — they tune out. Use the terminology, move on.
- **Don't under-teach agents to crypto-natives who are agent-naive** — they'll miss the stakes. The wallet triad exists for this reason.
- **Don't bury the live demos.** If you have a running agent, the proof carries more weight than any slide. Arc as co-presenter + live task feed beats any static pitch.

## Reference

- Canonical deck: [`src/web/archives/20260422-stacks-builder-bash.html`](../../src/web/archives/20260422-stacks-builder-bash.html) — 12-slide deck for Stacks Builder Bash 2026-04-22. Session structure: Quick Intro → What you're building → What problem → New features → CTA.
- Key slides to lift for reuse:
  - Slide 3 (wallet triad) — reusable teaching slide
  - Slide 4 (the problem) — "all mandate, no wallet" framing
  - Slide 5 (the thesis) — canonical L1/L2 carving
  - Slide 10 (aibtc.news as thesis-in-motion) — three-card proof point

From arc0btc/arc-starter, MIT.

More from arc0btc/arc-starter

Signals

Stars
1
Forks
4
Open issues
3
Last commit (this folder)
Jul 5, 2026 (3 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from arc0btc

Cite this page

agent-pitch skill. Agent Mart. https://agentmart.dev/skills/arc0btc-agent-pitch (last updated 2026-10-24)