An agent with a key, a trail on Bitcoin and two skills makes a short film about itself, runs a critic loop until it ships, and marks the result on the chain. Nothing is pasted in by hand; the agent reads two SKILL.md files and one brief. Done on 2026-10-19 on txbt4, the BLAKE2b testnet.
mkdir self && cd self
npm create agent@latest gitmark
An agent's life on the chain: init to exodus (gitmark workflow for LLM agents)
The page version, kept current, is https://git-mark.com/start.html. This gist is a copy.
A workflow for an LLM agent (or a person driving one). The agent gets a key, gets coins, anchors its commits to Bitcoin, announces the repo on Nostr, and when it is done hands the coin back and closes the trail. Every step is one command; every step has a check. Do the numbered steps in order and stop at the first check that fails. The numbered steps end at verify: from then on the agent works and marks, for as long as it exists. Retiring an agent is
| Verifying that I control the following Nostr public key: npub1melv683fw6n2mvhl5h6dhqd8mqfv3wmxnz4qph83ua4dk4006ezsrt5c24 |
Heads-up limit hold'em against a bot, for sats, where every hand settles the moment it ends and nobody takes a cut. Test coins only for now (txbt4, the BLAKE2b testnet4): worth nothing, which is the point while the rails are new.
flowchart LR
2026-08-11 · testnet4 · everything below is publicly verifiable
Yesterday a browser game and a dice house — separate sites, separate repos, zero communication between their servers — moved money between themselves correctly. The player carried it. A Bitcoin block rolled the dice. The whole history is signed by one key and notarized on-chain, and you can check every claim in this post yourself.
Seal your gold to your own key, prove it on Bitcoin, and stake it at the dockside dice — then bring your winnings home. Everything below is testnet: the coins have no value, and every move of your gold is signed by your key, in your browser. The server never holds your key.
Before you start: you need a nostr keypair (a 64-hex private key). Use a throwaway or testnet key, not your main nostr identity — the key is kept in your browser's localStorage for signing, which is fine for play and wrong for
This document evaluates BIP-110 using principles associated with IETF rough consensus and engineering review.
It does not assess any particular activation mechanism, signaling threshold, deployment schedule, or political campaign. Its scope is limited to the technical problem identified by BIP-110, the proposed restrictions, and the substantive objections that have been raised.
How a terminal habit became a realtime web dashboard — and accidentally became the first real JSS plugin, with a keypair for an identity and zero inotify watches.
For ages I've had this in a terminal:
Bitcoin is built on an uncomfortable truth: adversaries exist.
That is why we have proof of work. That is why nodes validate. That is why wallets check addresses. That is why signatures are verified, blocks are rejected, and invalid money is ignored.
Nobody says, “I do not think fake signatures will be a problem.”
Michael Saylor’s Strategy thesis begins with a simple provocation: cash is not a neutral asset. It looks safe because its nominal value is stable, but over long periods its purchasing power is deliberately eroded. Modern fiat money is designed around managed inflation. The Federal Reserve, for example, explicitly targets 2 percent inflation over the longer run, which means the monetary system treats mild currency debasement as policy, not as a bug. Bitcoin, by contrast, is built around a fixed supply thesis: no central issuer, no discretionary monetary committee, and a hard cap widely understood as 21 million coins.
This is the philosophical core of the Saylor view. Fiat cash is excellent for near-term liquidity, payroll, tax payments, and avoiding forced sales. But as a long-term store of value, it is structurally impaired. It is a melting ice cube. Bitcoin, in this framework, is not just a speculative asset. It i