meshbay.git/packages/meshbay-hub/tests/test_hub_api.py, branch main

fix(hub): present a node-audience token in the handshake, not the hub session token

2026-09-25T15:24:16Z

A member authenticated to a node in the MNP handshake with its hub *session* token — scope=user, valid at the hub API for hours. A node operator is in the threat model, so this handed them a live hub credential for the member: enough to enumerate the member's other groups, act as them, and (before the previous commit closed it) take the account over. The node genuinely needs a hub-signed membership assertion, so the fix is to make that a separate credential that opens nothing at the hub API. Two audiences signed by the one hub key (meshbay_common/tokens.py): - HUB_API_AUD — session tokens (login, device-auth, node-auth, refresh), used for hub calls and signaling. decode_access_token now binds this audience, so an MNP token cannot be replayed against the hub API. - MNP_AUD — a short-lived token a member presents to a node and nothing else, from POST /v1/nodes/mnp-token. authorize_token now binds this audience, so a session token presented to a node is refused. This closes the disclosure. The node's own self-decode (hub_client.py) reads its node token with audience=HUB_API_AUD. The client fetches the MNP token inside transport.connect() (and on every reconnect) using the session token, so callers are unchanged and signaling keeps using the session token. No regression to a long session: the MNP token is checked once, at the handshake, before any proof — a film already playing is not re-authenticated, so a 15-minute token does not interrupt a 4-hour film; reconnects refetch a fresh one. Denylist and membership checks are unchanged (the MNP token carries sub/jti/groups). Tests: authorize_token refuses a session/no-audience token and accepts an MNP token; the hub API refuses an MNP token; POST /v1/nodes/mnp-token is minted only for a member's own session. Verified red-before/green-after; common, node and hub suites green (the pre-existing test_cli_golden failure is an argparse/pytest prog artifact unrelated to this change). Still to do before deploy (B2): bump the MNP version and client.minimum so a stale desktop client is told to update rather than getting a handshake refusal, update docs/MESHBAY_DESIGN.md and MESHBAY_NODE_PROTOCOL.md, and validate against a real node locally, then deploy hub+node+SPA atomically. Co-Authored-By: Claude Opus 4.8

style: the 98 ruff could not fix, so the linter is a signal again

2026-09-19T12:39:38Z

The pass before this applied ruff's own fixes. These are the ones needing a decision, and the point of doing them is that `ruff check .` now passes: a linter reporting 98 known-acceptable findings reports nothing, because the next real one arrives invisible. **Lines over 100 (70).** Mostly wrapped where they stood. Two exceptions: the aligned trailing comments in `protocol.py`'s message table were shortened rather than wrapped, because wrapping one row of a table breaks the table; and in `models.py` the column comments moved above their columns for the same reason. **Imports below the first statement (14).** `csam.py` kept its FastAPI imports under a section header halfway down the file; two node tests had a constant and a `pytestmark` wedged between two import blocks. Moved, not suppressed. **Bindings nothing reads (4).** Three in tests, where the call stays and only the name goes — `_user(client, "listener")` is there to create the user, not to return one. The fourth was in `revocation.py` and was not a lint finding at all: `_connect_and_listen` opened an httpx stream to the WebSocket URL, did `pass`, and then opened the real connection through the `websockets` library. One pointless request per connect, left over from before that library was used directly. Removed, and `httpx` with it. **`l` as a name (4)**, **semicolons (6)** in the POC spikes, and the rest. 2893 passed, the same count as the two commits before it. `meshbay_node/revocation.py` is worth a decision separately: 154 lines that nothing imports, superseded by `hub_client.maintain_ws`'s `on_revocation`. This commit only stopped it failing the linter. Co-Authored-By: Claude Opus 5

style: ruff's own fixes, mechanically applied

2026-09-19T12:24:13Z

`ruff check .` had gone unrun long enough to report 568 errors, which is the same as having no linter: the next real finding would have been invisible in the noise. This is the 521 it fixes by itself, in 173 files, and nothing else — the 98 it cannot fix are the next commit. What actually changed: import sorting (225), imports nobody used (87, none of them a re-export — no `__init__.py` is touched, which was the one way this could have broken an import elsewhere), `datetime.timezone.utc` to `datetime.UTC` (69) and `asyncio.TimeoutError` to `TimeoutError` (18), both plain aliases on the 3.12 this project requires, `Optional[X]` to `X | None` (24), and f-strings with nothing to interpolate (19). Checked rather than assumed: every module in the three packages still imports, and the suite is 2893 passed — the same count, test for test, as the merge before it. Co-Authored-By: Claude Opus 5

fix(hub): the password verifier is Argon2id 64 MiB, and a hash's version names its parameters

2026-09-17T10:43:09Z

`pw_version` 4: Argon2id 64 MiB, t=3, lanes=4 — RFC 9106's second recommended setting. A v3 hash (256 MB) still verifies at its own parameters and is rewritten at the new ones on the next sign-in, through the rehash path that already existed. Why not more. The verifier matters against an offline attacker holding the database; online guessing is bounded by the sign-in lockout. That attacker pays the client's 600 000 PBKDF2-SHA512 iterations and the hub's Argon2id per guess, since `auth_key` is 256 bits and cannot be searched directly. Memory above 64 MiB multiplies that cost by a constant — at most 16 at 256 MB, less with PBKDF2 counted — while the hub pays the same memory at every sign-in, one derivation at a time. Measured on meshbay.org: 450 ms at 256 MB, 105 ms at 64 MiB, so a burst of sign-ins clears about four times faster. Changing the current version exposed a latent lockout. `hash_password` always used the current version's parameters, while the raw-password scheme recorded `pw_version = 2` — harmless while versions 2 and 3 shared their parameters, and with version 4 every legacy registration and v1→v2 rehash would have stored a 64 MiB hash labelled 256 MB, which nothing could then verify. Seventeen tests caught it. `hash_password` now takes the version it is hashing for. The OpenSSL deadlock between two concurrent `lanes=4` derivations is the same at 64 MiB, so Argon2 stays on its single worker. The loop-stall test measures against a v3 hash, because half of a 45 ms inline derivation is too close to scheduling noise to be a reliable bound. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01LcF3QKWii7uQ2kSyXErzCt

feat(hub): usernames are at least 8 characters at registration

2026-09-15T00:21:01Z

Existing shorter accounts keep signing in. Test usernames padded to match. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01XuNrwLf5EFWCMHzfoEvnpm

fix(hub): a node that hosts nothing is not a free target, and a private group's hosts are its members'

2026-09-13T19:18:23Z

Two checks that were not performed, in the same shape: something true of the public case, written so it did not apply to the rest. **Signaling.** The membership check read `if node_group_ids:`, so an empty set skipped it — membership, group status and the public-group gate together — and the offer was relayed. Since AV1 made an empty claim mean "no groups" rather than "all of my owner's", that is the *normal* registration of a node hosting nothing: the unconfigured node left running, the machine in this register's founding incident, and the one least able to absorb the work. Each offer makes it allocate an RTCPeerConnection and gather ICE, 30 a minute, which is H6 restored in exactly the case AV1 made common. It is refused now. Nothing legitimate is lost: such a node refuses the handshake anyway — `group_id` is mandatory (M1) and a node with no group key refuses (NS8) — so this declines work the node would decline one step later at its own expense. **The node list.** `GET /v1/groups/{id}/nodes` checked membership only for a public group with public groups switched off. A private one answered any authenticated account that knew the id — which an ex-member knows for ever — with the ids and public keys of the machines hosting it. §7.4 already stated the property for the public case. Membership is required now unless the group is public and public groups are on. Nothing needs the list before joining: an open join writes the membership row first, and an invitation registers the invitee's when the code is created, both checked before making the change. `test_webrtc_signaling_roundtrip` had to be repaired, and is the finding restated: it registered a node in `_connected_nodes` and never in `_node_groups`, so it drove the relay through the hole rather than through the door — the check it routed around was the defect. It now registers the group a real node registers. §7.2 and §7.3 gain the rules, the register AV24 and AV25, and test_availability_between_members.py two cases that fail on the property against the previous source. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01UMxEQadpzPkYLFf5CYKhpW

fix: bound what one member can cost the others

2026-09-12T14:36:54Z

An availability review, prompted by the group claim above: a participant supplies input — who else bears the cost? Six answers where the cost fell on someone other than the sender, and none of them needs an attacker. AV3 `chat_notify` carried a `group_id` the hub believed, so any connected node could write a notification to every member of any group on the hub, carrying a display string of its choosing, with its account having no relation to that group. This is the group claim again, two hundred lines further down the same socket. Gated on what the node is registered for, and metered: the fan-out is one write per member. The budget expires by time rather than on disconnect, or reconnecting would refill it and a node token is good for an hour. AV4 A swarm source named its own `endpoint` as free text documented as "ip:port", so an account could publish a third party's address — H6's `peer_ip` defect, never applied here. Nothing dials a swarm source today, which is the only reason it was not already a reflection primitive. It is a transport and a port now, never a host, and the number of hashes one account may claim is bounded: rows were keyed (hash, account) with no cap at all. AV5 `handle_webrtc_answer` resolved any pending `peer_id` from any node's socket. The answer is the SDP a browser then connects to. That this had not happened rested on a uuid4 being unguessable. AV6 `relay_register` had no authentication of any kind: it compared `pk_relay` against the approved value, which is a *public* key, so anyone who could read it could rewrite where the hub tells nodes to send relayed traffic. The module docstring promised signed JWTs and `jwt` was imported and never used. AV7 The node held unlimited peer connections and kept one that never completed a handshake for the life of the daemon. H6 bounded what one unauthenticated peer costs; the hub's cap is three offers in flight per *account*, a limit on each caller and not on the machine, so an operator's exposure grew with the size of their groups. AV8 `invite-notify` put a request-supplied `group_name` into the subject of an email the hub sends under its own domain, to any account, with no rate limit. The name comes from the group row now. The tests are two accounts each, in one file that says why: a one-member test proves a one-member property, and every finding here needed a second person to exist at all. Each was checked against the unfixed code. Two did not survive that check and were rewritten — one re-enacted the disconnect path instead of running it (hence `forget_node`), the other called the reaper itself and would have passed with the call removed from `handle_offer`. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01T4YmK41VsEURWFdop4EEeT

fix(hub): serve the SPA under a fingerprint of what it is

2026-08-16T18:57:52Z

`Cache-Control: no-cache` requires a browser to revalidate, but it only binds one that asks. A browser that cached app.js before that header existed applies heuristic freshness instead — a fraction of the file's age, which for a file dated weeks ago is days — and never asks. It then runs an old player against a new node. That cost most of a session. A phone kept a player without the read-ahead bound and filled the browser's buffer ceiling at 106 MB, the exact symptom the bound had been written to remove, for an hour after the bounded player went live. A fix that is written, tested, deployed and served, and still not what runs, is indistinguishable from a fix that does not work. The whole module graph now lives under `/a//`. A path prefix rather than a query string, because relative imports inherit it: `app.js` reaching for `./i18n.js` gets the build it was written against, and never a mixture of two — which does not render a stale page, it fails to link. The URL changes with the content, so those may be cached hard. `sw.js` stays at the root. Its scope is its own path, and under the prefix it would no longer control the pages whose downloads it exists to intercept.

feat(hub): leaving a group, a cap on public ones, and hosting as a precondition

2026-08-16T13:28:28Z

Leaving is its own endpoint rather than a relaxation of the owner's removal check — an authorization rule with an exception in it is the one that gets read wrong later. The owner cannot leave: the group would be left with nobody able to admit, edit or delete it, which is the answer removal and account deletion already give. Public groups are capped at ten live ones per owner. They are the ones that cost other people something — listed in Discover, joinable by anyone — so a script that opens hundreds fills the directory for everybody. Private groups are invisible to non-members and are not capped. Hub staff are exempt; the cap is anti-spam, not a rule about running an instance. Creation is the only place it can be checked, and deliberately so, because PATCH refuses to change visibility at all. A group is now listed only once a node has announced that it hosts it. Before that it has no files, no key and nothing to connect to, so showing it to a member produces a name they cannot open and cannot be told why; its owner still sees it while they set the node up. `meshbay-hub prune-groups` collects the ones that never got a node, meant for cron, with --dry-run. The migration backfills hosted_at from created_at: without that the first run would have deleted every live group. Presence rides on the group list itself, read from the signaling registry the hub already keeps — no poll, no timer. It says a node is connected *to the hub*, which is not a promise that this browser can reach it and not something a dishonest hub could not fake; the client downgrades it on a connection it tried and failed, which is the evidence that concerns the reader. Co-Authored-By: Claude Opus 5

feat!: identity keys per node — C4's blast radius drops to one operator

2026-08-14T15:51:48Z

One keypair was copied to every node its owner joined, so cracking the bundle on any single node yielded the identity used on all of them: their content on other operators' machines, and the ability to sign as them anywhere. That lateral reach was the part of C4 worth attacking. Each node now gets its own keypair, generated the first time its owner joins it and left with that node alone. An operator who cracks what sits on their own disk holds a key that is a stranger to every other node — and on their own node, one that unlocks nothing they did not already hold: they serve the content, the index and every byte of it by design. Nothing changes for the user. A first contact with a node already needed that operator's code, and the key is created in the same step; a second browser still recovers it from the node with the passphrase alone. Two operators can also no longer tell they host the same person by comparing keys. BREAKING, and deliberately without a compatibility path — the deployment is wiped for the next demo: - users.pk_ed25519 / pk_x25519 dropped (migration a7c31f9e40b2) - registration no longer sends or stores a key - PUT /v1/users/me/keys and regenerateKeys() gone; rotation is now `member unpin` plus a fresh code, decided on the machine that pinned it - /pubkeys returns an account id and a node's linking key. It was the directory H3 read, and nothing wraps for it any more - the pk_user JWT claim is gone That last one closed a live defect the inventory turned up: the node recorded pk_user as the uploader's identity and authorized deletion against it, so a hub issuing a token naming its own key could delete anyone's uploads on any node. Attribution now uses the key the node itself pinned. A simplification falls out. Registration generates nothing, so a scripted signup is a real account: `demo.py bootstrap` takes a wiped hub and node to a working demo with no browser, which was impossible while keys were born in one. Also fixes, found by running it on a wiped deployment: the key handed back on a join now belongs to the group the connection is for, not the group named in the invitation — an operator pairs node-wide but redeems the code while opening a group, and expects to read it. Tests: 343, including the two that state the property — a key pinned by one node is refused at another, and someone else's code does not admit it. Verified end to end against a wiped hub and node: bootstrap, pair, invite, join, download, stream, second browser, revoke. Design: docs/per-node-identity-v1.md Co-Authored-By: Claude Opus 5