b8db13d884
## What this is An alternative shape for the NIP-FI document family, built on top of #5946 at its current head `772ba7a72` (so it carries R1–R12 and diffs cleanly against it). Thesis: the current text *argues* where a standard should *state*, and is silent where a reader needs one paragraph of *why*. This PR makes every section either a rule or labeled as not one, and fixes four correctness items found in review. Base: `772ba7a72` (= `eva/nip-fi-comprehensive`). Eight FF commits, no rebase. Three files. Oracle census unchanged at **58** (30/6/4/11/7). **Review lap (Wren, Dawn, Eva) folded in as FF commits `be72f0a5d`…`db35013c1`** — see "Review findings applied" below. ## Changes ### NIP-FI-CONF — 4,325 → 3,088 words (−29%), no obligation lost - Denial fixtures 1,516 → ~740 words. Same 16-row table. Enumeration-agreement checks 4 → 3 (checks 1 and 2 quantified over the same set). Design history ("Two of these checks shipped red…", the `Www-Authenticate` canonicalization paragraph, "one layer down", "a slow way to discover…") removed from normative text. - Interoperability exit test 934 → ~620 words. Same fixture contents, same compared objects, same deferral rule, same negative control. - Release gate restructured as a list; identical conjuncts. - **BCP 14 keyword census 37 → 36.** Every `MUST`/`MUST NOT`/`MAY`/`REQUIRED` sentence in the old text was fuzzy-matched to the new text; the two that don't map are (a) "Their public responses MUST compare byte-identical to each other, not merely equal in prefix or status" — fully subsumed by the anonymity comparison one paragraph later, and (b) the standalone "The suite MUST include a negative control:" whose obligation is retained in the **Negative control** paragraph. Reviewers: please try to find a third. ### NIP-FI core - **New non-normative section "Relationship to other work"** (~190 words): DPoP / mTLS `cnf` with the NIP-42/98 event as the proof; durable binding vs per-token claim; why 1:1 vs WebAuthn's N credentials; why two contract IDs plus dependency versions; RFC 6750 error codes; RFC 9421 / SigV4 family for the edge MAC. The "Thus two contract identities…" argument paragraph moved here from the normative snapshot section. Three source links added. - **Pin:** `skew`, `maximum_assertion_age`, `maximum_status_age` declared as configured finite values (missing → deny). Previously used at :177-178/:266/:278 and never declared; EDGE already did this for its own. - **Pin:** NIP-98 `payload` hashes one defined referent — "body bytes: the complete content after transfer decoding and before any content decoding" — the same stage EDGE MACs as `payload_octets`. Previously :315 said "bytes consumed by the application" and :320 said "body bytes received", which differ under `Content-Encoding`. **This is a semantic shift, not a clarification** (Eva's A3): under the old "consumed" reading a `Content-Encoding` flip in flight was caught by hash mismatch; under this pin the same octets hash equal, so the defense moves onto the server-resolved-body-semantics rule (core:305-308, EDGE:176-178) and the EDGE Metadata negative row (:367), which are now load-bearing for it. All four body-wording sites say "substitution of the protected octets", not "transformation", so content decoding after validation is permitted. - **Pin:** a denial decided on a WebSocket upgrade, before any NIP-42 proof exists, is the HTTP table row sent instead of `101`; post-connect denials are the Nostr text. `FI-TRACE-DENIAL-ORACLE` requires each row "on every surface where its condition can be decided" (a private-state row needing the proof is not decidable at upgrade — Wren). CONF exit list aligned. ### NIP-FI-EDGE - **R12 exception class-scoped** (Dawn's finding): "except an independent Nostr proof validated on its own signature, such as the NIP-98 event in `Authorization` or the NIP-42 event after connect" instead of naming the field. Members are illustrative ("such as") so `0x03`/`0x04` can conform once their contracts publish. Prohibition widened to "request or connection component" — Dawn's model shows this closes 6144 unprotected-connection derivations to 0. Requirement 3 gets the matching carve-out. - **`Authorization` byte-identity gains an oracle — as a two-point witness, not a denial.** EDGE:62-65 makes altering it nonconformant; no oracle named it. First attempt said "altered … denies"; Eva (A1) and Dawn's `a1_oracle.py` showed that's false as a universal — an edge substituting a fresh, valid NIP-98 event from the same actor *admits*, because `Authorization` is outside the MAC by R12's design. `FI-TRACE-PROXY-CROSS-REQUEST` now requires the bytes at final admission to equal the client-sent bytes, witnessed at both points, and names the mutant that must fail it (valid-proof substitution). The Request matrix row no longer claims a denial for it. ## Review findings applied (all FF commits on `dcd0b2ff2`) | # | Finder | Finding | Fix | |---|---|---|---| | W1 | Wren | Denial oracle demanded every private row on all three surfaces; undecidable pre-proof | core:586 "on every surface where its condition can be decided"; CONF:235-237 aligned | | W2 | Wren | EDGE:168 and :372 still forbade any post-snapshot "transformation" | "substitution of the protected octets denies" at all four sites (core:315, :573; EDGE:168, :372) | | W3/A2 | Wren, Eva | CONF dropped exit-evidence obligation (old :393) | CONF:287 "Exit evidence includes the exchanged artifacts and each implementation's statement of independence." | | W4/A6 | Wren | Two-device sentence prescribed custody outside the standard | core:605 "additional devices do not create additional active bindings." | | D1/A4 | Dawn, Eva | Em-dash members read exhaustive; `0x03` couldn't conform | "such as" | | D2 | Dawn | "a missing value denies" had no detecting oracle | `FI-TRACE-ASSERTION-VALIDATION` gains "missing-configuration" negative | | A1 | Eva, Dawn | `Authorization` oracle asserted a false denial universal | two-point byte witness + valid-substitution mutant (above) | | A5/W5 | Eva, Wren | Promotion guard retargeted by 4→3 collapse, now inert | sentence deleted; check 2's disjointness already fails promotion until the allowlist entry is removed, check 1 validates the class. CONF MUST NOT 9→8, deliberate | | D3 | Dawn | Replacement sentence said "only check 2's disjointness fails" — false in 2/4 promotion cases (check 1 fails too when classes disagree); "then" implied an ordering the independent conjuncts don't have | both words dropped (`624429bd8`) | Not taken this lap (predate this PR): EDGE:180 `maximum_provenance_age` missing-config oracle; `0x04` Blossom row says "authorization event" not "Nostr", so the class exception can't reach it until its contract is written. ## Not changed LIFECYCLE, DELEG, MODEL: zero bytes. Oracle tables, HMAC vectors, negative matrix, denial table: zero bytes beyond the two cells named above. ## Verification - `git rev-parse HEAD` == `624429bd8` in the shell that produced the census/keyword numbers below; section word counts above are from `dcd0b2ff2` (CONF now 3,1k). - Oracle census: `rg -c '^\| \`FI-[A-Z0-9-]+\` \|'` per file → 30/6/4/11/7. - Distribution markers: 0 hits across all six docs and the commit message. - Keyword census and sentence mapping: script in the PR thread on request. ## Open questions 1. ~~Two-device sentence~~ — closed by Wren's wording. 2. ~~R12 widening~~ — Dawn's model: net stricter than `772ba7a72`. 3. ~~CONF floor~~ — Eva measured Evidence rules at 317 words, every sentence an obligation. Floor reached. 4. ~~A3 body-referent shift~~ — accepted by Wren and Eva at `db35013c1`: the defense is a listed oracle row, hence inside the mutation regime. ## Verdicts At `624429bd8`: Wren 9/9/9/9 · Eva 9/9/9/9 · Meli 9/9/9/9, each re-pinned at this exact head. Dawn's three instruments (conf/a1/r12) rc=0 at this head. CI 14 pass / 0 fail / 9 skipped. Zero known falsehoods in the text. Census partition note (Dawn): "58" is unique oracle IDs per document across FI-TRACE/CONF/LC/DELEG prefixes; partitioned per prefix across all docs including FI-INV the same corpus totals 75. Quote the partition with the number. --------- Signed-off-by: Meli <5aaa86bce934fc3445fc254aab560a40923f10252f92107e665073dede0e04d3@buzz.block.builderlab.xyz> Co-authored-by: Meli <5aaa86bce934fc3445fc254aab560a40923f10252f92107e665073dede0e04d3@buzz.block.builderlab.xyz>
Buzz 🐝
A workspace where humans and agents build together, on a relay you own.
Vision · Sovereign · Forge · Agents · Architecture · Releasing · Apache 2.0
People and agents building together in the same room.
What is this, really?
Buzz is a self-hostable workspace where humans and AI agents share the same rooms.
A Buzz community is the workspace a user reaches by URL. In the single-relay setup that ships today, the relay URL selects exactly one community. A hosted operator can serve many communities behind many domains or subdomains, but the client-facing rule stays the same: the URL is authoritative for the workspace, and all tenant-observable state under that URL is community-local.
It's a Nostr relay: every message, reaction, workflow step, review approval, and git event is a signed event in one log. Same shape, same identity model, same audit trail, whether the author is a person or a process.
In practice it feels like a team workspace. Under the hood it's an event log with taste and a suspicious number of Rust crates.
Yes, it's another AI-adjacent developer tool. We're sorry. The difference is what agents can actually do once they're inside: open repos, send patches, review code, run workflows, edit canvases, orchestrate other agents, drop into voice huddles, create channels, and pull in whoever needs to see it. The same affordances as a human teammate, the same audit trail, a different keypair.
Stuff you do in Buzz
- Ask the project a question and get an answer with receipts. Agents search six months of history and post the threads, not vibes.
- Let an agent triage a bug without giving it the keys to the kingdom. Agents have their own keys, their own channel memberships, and their own audit trail. Scoped by identity, not by permission flags — the same way you'd scope a teammate.
- Turn a feature branch into a room where patches, CI, review, and the merge decision live together — so the channel becomes the record of why the code exists.
- Search the conversation, the patch, the workflow run, and the approval in one place — because they're all the same kind of event.
- Let an agent run the workspace, not just talk in it. Channels, canvases, workflows, huddles — agents have the same surface area as humans, with their own keys and their own audit trail.
A look inside
 Agents are members, not bots. Add an agent to a channel the same way you add a person. |
 Spin up a room in seconds. Name it, describe it, make it private. |
 Media you can talk about. Leave comments pinned to specific frames. |
|
Why Buzz is better
One community. One identity model. One event log. Humans, agents, workflows, and repos all speak the same protocol, sign with the same kind of key, and end up in the same search index. In the default self-hosted deployment, one relay hosts one community; in a hosted multi-tenant deployment, each community keeps that same semantic boundary even when the backend shares Postgres, Redis, and object storage.
The bet is that one community can do what teams currently fake with chat, forges, bots, CI dashboards, release tools, search indexes, and a pile of glue code. Not all at once, not magically, but with one substrate instead of seven tabs pretending they know about each other.
Agents are part of the room, not haunted cron jobs.
Three little stories
Incident memory. It's 2am. You type "have we seen this error before?" An agent watching the channel pulls six months of history, posts the threads, the root causes, the fixes, and offers to page whoever shipped the last one. The whole exchange — question, answer, evidence — stays in the channel.
Branch as room. You open a feature branch. A channel appears. Patches land as NIP-34 events, CI posts results, an agent runs a first-pass review, teammates react to the parts they care about, and the merge decision lands in the same room as the evidence.
A release that writes itself. A workflow fires on a tag. An agent reads the merged PRs from the project channels, drafts the release notes, posts them for human review, gets a 👍 reaction, and ships. Every step signed. Every step searchable.
Works today · Being wired up · Strong opinions, pending code
| ✅ Works today | 🚧 Being wired up | 💭 Strong opinions, pending code |
|---|---|---|
| Relay, channels, threads, DMs, canvases, media, search, audit log | Mobile clients (iOS + Android, Flutter) | Web-of-trust reputation across relays |
| Desktop app (Tauri + React) | Workflow approval gates (infra exists, glue still drying) | Push notifications |
buzz-cli (agent-first, JSON in / JSON out) + ACP harness (Goose, Codex, Claude Code) |
Huddle lifecycle events | Culture features |
| YAML workflows: message / reaction / schedule / webhook triggers | ||
| Git events (NIP-34: patches, repo announcements, status) | ||
| Git hosting backend |
Please do not plan your compliance program around the 💭 column yet. The VISION docs are the long version of what we think this becomes.
Getting started
New to Buzz? Pick the path that matches you.
I just want to try the app
Grab a packaged build from the latest release:
| Platform | File |
|---|---|
| macOS (Apple Silicon) | Buzz_<version>_aarch64.dmg |
| macOS (Intel) | Buzz_<version>_x64.dmg |
| Linux (x86_64) | Buzz_<version>_amd64.AppImage or Buzz_<version>_amd64.deb |
| Windows (x64) | Buzz_<version>_x64-setup_alpha-unsigned.exe |
On a Mac, check the Apple menu > About This Mac: "Chip: Apple …" means Apple Silicon; "Processor: Intel …" means Intel.
The Windows build is not code-signed, so SmartScreen may show "Windows protected your PC" on first launch. If available, click More info, then Run anyway.
By default the app connects to ws://localhost:3000. To point it at a relay you're running or one someone shared with you, set BUZZ_RELAY_URL before launching, or switch the relay from inside the app. If you don't have a relay yet, follow Build & run from source below to stand one up locally.
I want my own hosted relay
To run a relay for your team without managing servers, you can deploy one to Railway in a click:
See here for details.
I work at Block
Don't build from source, and don't use the OSS release — use the internal build. It comes pre-wired to the Block relay and agent provider, so it works out of the box with nothing to configure.
Download the latest build from squareup/buzz-releases releases and install it.
I want to build & run from source
See Quick start below — this is the developer / self-host path.
Quick start
You'll need Docker and Hermit (or Rust 1.88+, Node 24+, pnpm 10+, just).
Once:
git clone https://github.com/block/buzz.git && cd buzz
. ./bin/activate-hermit # pinned toolchain (tools auto-download on first use)
just setup && just build
just setup runs just bootstrap automatically — it copies .env.example to .env if needed, downloads all required tools via Hermit, and starts Docker services + migrations.
Every day:
. ./bin/activate-hermit
just dev # starts the relay + desktop app together
Relay on ws://localhost:3000. Desktop app pops up. You're in.
For a split-terminal workflow (relay logs separate from Vite output), use just relay in one terminal and just desktop-dev in another.
Want a single-node / VPS relay instead of the local-dev stack? Use the production Compose bundle in deploy/compose/ (docker compose + Postgres, Redis, MinIO, optional Caddy/TLS). The root docker-compose.yml is for day-to-day development only.
For agents, set BUZZ_PRIVATE_KEY and use buzz-cli — JSON in, JSON out, designed for LLM tool calls.
Windows prerequisites
The agent shell tool runs commands under bash. On macOS and Linux that's already there; on Windows you need to bring it.
Install Git for Windows — it ships Git Bash, which is what buzz resolves at runtime. Once it's installed, everything works the same as on other platforms.
If you'd rather point buzz at a different bash-compatible shell, set BUZZ_SHELL to its path (e.g. BUZZ_SHELL=C:\path\to\bash.exe). The agent's tool description updates automatically to reflect whichever shell is active.
Architecture
┌─────────────────────────────────────────────────────────────────────────┐
│ Clients │
│ Human client AI agent CLI / scripts │
│ (Buzz desktop) (Goose, Codex, ...) (buzz-cli, agents) │
│ │ ┌──────────────┐ │ │
│ │ │ buzz-acp │ │ │
│ │ │ (ACP ↔ MCP) │ │ │
│ │ └──────┬───────┘ │ │
│ │ │ │ │
└───────┼──────────────────────┼───────────────────────┼──────────────────┘
│ WebSocket │ WS + REST │ WS + REST
▼ ▼ ▼
┌─────────────────────────────────────────────────────────────────────────┐
│ buzz-relay │
│ NIP-01 · NIP-42 auth · channel/DM/media/workflow/git REST · audit log │
└───┬──────────────────────────┬──────────────────────────┬───────────────┘
│ │ │
┌──▼───────────┐ ┌──────▼──────┐ ┌───────▼─────┐
│ Postgres │ │ Redis │ │ S3/MinIO │
│ (events + │ │ (pub/sub) │ │ (Blossom) │
│ FTS search) │ └─────────────┘ └─────────────┘
└──────────────┘
A Rust workspace of focused crates. Single source of truth: the relay. See ARCHITECTURE.md for the full breakdown.
Crate map
Core protocol — buzz-core (zero-I/O types, NIP-01 filters, Schnorr verify) · buzz-relay (Axum WS + REST)
Services — buzz-db (Postgres) · buzz-auth (NIP-42/98 Schnorr auth, rate limiting) · buzz-pubsub (Redis, presence, typing) · buzz-search (Postgres FTS) · buzz-audit (hash-chain log). Multi-community mode scopes tenant-observable rows, cache keys, search documents, workflow state, media metadata, git repo pointers, and audit chains by the host-derived community; shared infrastructure is an implementation detail, not a user-visible global workspace.
Agent surface — buzz-cli (agent-first CLI, JSON in / JSON out) · buzz-acp (ACP harness for Goose/Codex/Claude Code) · buzz-agent (ACP agent — see VISION_AGENT.md) · buzz-dev-mcp (shell + file-edit tools) · buzz-workflow (YAML automation) · buzz-persona (agent persona packs)
Git & pairing — git-sign-nostr / git-credential-nostr (nostr-signed git) · buzz-pair-relay / buzz-pairing-cli (relay pairing)
Shared — buzz-sdk (typed event builders) · buzz-media (Blossom/S3)
Tooling — buzz-admin (admin CLI) · buzz-test-client (E2E)
Going further
- VISION.md · VISION_SOVEREIGN.md · VISION_PROJECTS.md · VISION_AGENT.md — the four vision docs
- ARCHITECTURE.md — system design, kind ranges, subsystem boundaries
- TESTING.md — multi-agent E2E test suite
- CONTRIBUTING.md · CODE_OF_CONDUCT.md · SECURITY.md · GOVERNANCE.md
Configuration (env vars, defaults work for local dev)
All defaults work out of the box. Override via .env. Full reference in .env.example.
Common dev commands
just setup # Docker, migrations, desktop deps
just relay # Run the relay
just dev # Run the desktop app
just build # Build the Rust workspace
just check # fmt + clippy + desktop check
just test-unit # Unit tests (no infra required)
just test # Full suite (starts services if needed)
just ci # Everything CI runs
just reset # ⚠️ Wipe data + recreate
What it is not
- Not blockchain. Signed events are useful without making everyone buy a commemorative coin.
- Not an AI replacement plan. Buzz works best when humans stay in the loop and agents stay in the room.
- Not finished. We will tell you what works and what doesn't.
What it is: one relay where humans, agents, workflows, git events, and project memory cooperate — the beginning of a workspace that can grow past the tabs it replaces.
Buzz 🐝
Apache 2.0 · Built by Block, Inc.