logicsrc/packages/opencreds/src/items.ts
Anthony Ettinger 80a36269bb
Some checks failed
CI / build (push) Has been cancelled
test / test (push) Has been cancelled
Add the LogicSRC OpenCreds specification (#140)
* Add the LogicSRC OpenCreds specification

Leaving a password manager means writing every secret you own to disk in
the clear, and losing whatever the spreadsheet had no column for. A CSV
is plaintext by construction, lossy by omission, and carries no
integrity: nothing in it says which rows were meant to be there, so a
truncated import looks exactly like a complete one.

The same gap showed up inside LogicSRC. `logicsrc credentials` moves
.env secrets and SSH keys through end-to-end-encrypted team vaults, but
it can only model a key/value pair. A card, a passport, a login with a
TOTP seed, or an OAuth account with a refresh token are all things
people already keep in a vault, and none of them are a key/value pair.

OpenCreds defines three things: the item, the vault, and the database.

- Six item types (login, card, identity, note, key, account) as one
  record with a type and a named field group, so everything the user
  typed lives in a single encrypted blob. Codes 1-4 match MarkSyncr's
  deployed vault and are not renumbered; compatibility is cheaper than
  elegance.
- AES-256-GCM over that record with the item id bound as AAD. Without
  it, anyone with storage write access could move a low-value login's
  ciphertext into a high-value row and watch what the user does next.
- A key hierarchy where the user key is random, not derived, so a
  password change re-wraps 32 bytes rather than re-encrypting a vault.
  The auth hash comes out of a different HKDF label than the wrap key,
  which is what lets it reach a server at all.
- A portable .opencreds file, encrypted by default, whose header is the
  AAD over the payload -- so the manifest is authenticated by the same
  tag as the data and a truncated import fails rather than reporting
  success. The plaintext form exists because people move to products
  that read nothing else; it is opt-in, confirmed, 0600, and labelled
  "protected": false in its own header.

Namespaces are carried as data, not fixed by the spec: labels are
compiled into every ciphertext a vault has written, so editing one does
not migrate a vault, it makes it undecryptable. MarkSyncr's deployed
vault is conformant by declaring `marksyncr`.

Ships: prd/0004, nine spec pages under docs/opencreds/, six JSON
Schemas, the @logicsrc/opencreds reference implementation with CSV
importers for five products, `logicsrc vault` and the standalone
`opencreds` binary, and the spec page at logicsrc.com/opencreds.

`vault` rather than `creds` because `creds` is already an alias of
`logicsrc credentials`, and the two are different: one moves a pair
between providers, the other stores a record.

@logicsrc/validators now registers every schema by $id before
compiling, so the database schema can $ref the item and manifest
schemas rather than restating them.

120 tests, including CLI end-to-end coverage of the masking rules,
exit codes, and the manifest-mismatch path.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QRQrfuwuYKKV5UB9kLHuX5

* Make the OpenCreds conformance claim executable

The conformance page described a fixture suite and an `opencreds
conformance` command that did not exist. A specification that documents
a conformance surface it cannot run is a specification nobody can hold
to, including us.

`opencreds conformance` now runs the requirement list as code -- one
check per C-number, carrying its own id and level -- and emits the
report shape the spec publishes. It exits 2 when a MUST does not pass,
so it can gate CI directly. The reference implementation reports 29
passed, 0 failed, 1 skipped; the skip is C19, because key management for
the team profile lives in @logicsrc/plugin-credential-sharing rather
than in this package, and a skipped MAY does not affect conformance.

Fixtures are generated (`--emit-fixtures <dir>`) rather than
hand-written. A vector produced by an implementation and then verified
by it is worth more than a JSON file someone typed: the typed file
drifts silently when the format moves, and the generated one cannot.
Fourteen files, including an invalid/ set every conforming reader must
reject -- a wrong field group, a weak KDF, an unregistered namespace, a
short payload and a tampered manifest.

The CLI requirements stay with the end-to-end tests that drive the real
binary through a child process; a command cannot meaningfully check its
own exit codes, and a masked value that is only masked in the library is
not masked.

conformance.md and cli.md now describe what ships.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QRQrfuwuYKKV5UB9kLHuX5

* Add @logicsrc/opencreds to the lockfile

`npm ci` refuses a lockfile that does not match package.json, and the
new workspace package plus the CLI's dependency on it were never
recorded: the worktree was bootstrapped by hardlinking node_modules
rather than installing, so npm was never asked to update the lock.

Adds the workspace link and the package entry. No dependency versions
move.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QRQrfuwuYKKV5UB9kLHuX5

* Register PRD 0004, and stop the fixtures looking like real secrets

Two CI failures, both mine.

`prd/README.md` is generated by `logicsrc prd index --write` and the
scaffold test asserts it is current, so adding a PRD without
regenerating it leaves the repo's own conformance check failing.
Regenerated. The MCP test asserts the next free PRD id against the live
prd/ directory — its comment says it advances with every PRD added — so
it moves to 0005.

ThreatCrush flagged three of the example strings: a PEM header in the
item-model docs and in the conformance fixture, and an `sk_live_`
prefixed token. All placeholders, none real, but the finding is the
scanner working. A fixture only has to exercise the field, and a
real-looking private key header or live-key prefix sitting in the tree
trains both the scanner and the people reading its output to shrug at
exactly the shape that matters. Replaced with obvious placeholders
rather than suppressing the rule.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QRQrfuwuYKKV5UB9kLHuX5

---------

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-29 04:11:38 -07:00

340 lines
11 KiB
TypeScript

/**
* The item: creating one, editing one, and putting it in an envelope.
*
* Logins, cards, identities, notes, keys and accounts are not six features —
* they are one record with a `type` and a named field group. Everything the
* user typed lives inside a single encrypted blob, which is what makes password
* history free: it is an array in that blob, encrypted by construction rather
* than needing its own protected table.
*/
import { aesGcmDecrypt, aesGcmEncrypt, fromBase64, toBase64, utf8Decode, utf8Encode, uuid } from "./primitives.js";
import { itemLabel } from "./kdf.js";
import {
DEFAULT_NAMESPACE,
ITEM_SCHEMA_VERSION,
ITEM_TYPE,
MAX_HISTORY_ENTRIES,
type AccountGroup,
type CardGroup,
type Envelope,
type IdentityGroup,
type Item,
type ItemTypeName,
type KeyGroup,
type LoginGroup,
type Namespace,
} from "./types.js";
/** Empty field groups, so every item has a predictable shape. */
const EMPTY_FIELDS = Object.freeze({
login: (): LoginGroup => ({ username: "", password: "", totp: "", uris: [] }),
card: (): CardGroup => ({ cardholderName: "", brand: "", number: "", expMonth: "", expYear: "", code: "" }),
identity: (): IdentityGroup => ({
title: "",
firstName: "",
middleName: "",
lastName: "",
username: "",
company: "",
email: "",
phone: "",
address1: "",
address2: "",
address3: "",
city: "",
state: "",
postalCode: "",
country: "",
ssn: "",
passportNumber: "",
licenseNumber: "",
}),
note: (): Record<string, never> => ({}),
key: (): KeyGroup => ({
keyType: "",
algorithm: "",
publicKey: "",
privateKey: "",
passphrase: "",
fingerprint: "",
value: "",
path: "",
mode: "",
expiresAt: "",
}),
account: (): AccountGroup => ({
provider: "",
accountId: "",
handle: "",
email: "",
accessToken: "",
refreshToken: "",
tokenType: "",
scopes: [],
expiresAt: "",
environment: "",
}),
});
/** The group names, so a wrong-group check does not have to hard-code them twice. */
const GROUP_NAMES: readonly ItemTypeName[] = Object.keys(EMPTY_FIELDS) as ItemTypeName[];
export function isItemType(value: unknown): value is ItemTypeName {
return typeof value === "string" && value in ITEM_TYPE;
}
/**
* Create an item.
*
* The id is generated here, on the client, because it is bound into the
* ciphertext as additional authenticated data. Storage records this id rather
* than assigning one.
*/
export function createItem(type: ItemTypeName, fields: Partial<Item> = {}): Item {
if (!isItemType(type)) throw new Error(`Unknown item type: ${type}`);
const now = new Date().toISOString();
const group = EMPTY_FIELDS[type]();
const incoming = (fields as Record<string, unknown>)[type];
const item: Item = {
v: ITEM_SCHEMA_VERSION,
id: uuid(),
type,
name: "",
favorite: false,
folderId: null,
notes: "",
history: [],
createdAt: now,
updatedAt: now,
...stripGroups(fields),
};
if (type !== "note") {
(item as Record<string, unknown>)[type] = {
...group,
...(typeof incoming === "object" && incoming !== null ? incoming : {}),
};
}
return item;
}
/**
* Copy the top-level fields of a partial item, minus every per-type group and
* the fields this function owns. Unknown keys pass through: an item written by
* a later version must survive a round trip here.
*/
function stripGroups(fields: Partial<Item>): Record<string, unknown> {
const out: Record<string, unknown> = {};
for (const [key, value] of Object.entries(fields)) {
if (GROUP_NAMES.includes(key as ItemTypeName)) continue;
if (key === "v" || key === "id") continue;
out[key] = value;
}
return out;
}
/** "an account", "a login". These strings are read by people. */
function article(word: string): string {
return /^[aeiou]/i.test(word) ? "an" : "a";
}
/**
* Reject an item carrying a group that is not its own type's.
*
* A `card` group on a `login` item is either a broken importer or an attempt to
* smuggle a field past a type-based permission check; neither should be stored.
*/
export function assertGroupsMatchType(item: Item): void {
for (const name of GROUP_NAMES) {
if (name === "note") continue;
if (name !== item.type && (item as Record<string, unknown>)[name] !== undefined) {
throw new Error(`A ${item.type} item must not carry ${article(name)} ${name} field group`);
}
}
}
/**
* Record a password change in the item's own history.
*
* Called before overwriting the password, so the value being replaced is what
* gets kept. Returns a new item; does not mutate.
*/
export function recordPasswordChange(item: Item, newPassword: string): Item {
if (item.type !== "login") throw new Error("Only logins have password history");
const previous = item.login?.password ?? "";
const history =
previous && previous !== newPassword
? [{ password: previous, changedAt: new Date().toISOString() }, ...(item.history ?? [])]
: [...(item.history ?? [])];
return {
...item,
login: { ...(item.login ?? EMPTY_FIELDS.login()), password: newPassword },
history: history.slice(0, MAX_HISTORY_ENTRIES),
updatedAt: new Date().toISOString(),
};
}
/** Apply a partial update, merging the field group rather than replacing it. */
export function updateItem(item: Item, patch: Partial<Item>): Item {
const group = (patch as Record<string, unknown>)[item.type];
const next: Item = {
...item,
...stripGroups(patch),
updatedAt: new Date().toISOString(),
};
if (group && typeof group === "object" && item.type !== "note") {
(next as Record<string, unknown>)[item.type] = {
...((item as Record<string, unknown>)[item.type] as object),
...group,
};
}
if (next.history && next.history.length > MAX_HISTORY_ENTRIES) {
next.history = next.history.slice(0, MAX_HISTORY_ENTRIES);
}
assertGroupsMatchType(next);
return next;
}
/**
* The additional authenticated data bound to an item's ciphertext.
*
* Binding the id means a ciphertext cannot be moved from one row to another
* without decryption failing — without it, anyone with database write access
* could swap the ciphertext of a low-value login into a high-value one and
* watch what the user does next.
*/
function itemAad(namespace: Namespace, id: string, version: number): Uint8Array {
return utf8Encode(itemLabel(namespace, version, id));
}
export async function encryptItem(
userKey: Uint8Array,
item: Item,
namespace: Namespace = DEFAULT_NAMESPACE,
): Promise<Envelope> {
if (!item?.id) throw new Error("An item must have an id before it can be encrypted");
if (!isItemType(item.type)) throw new Error(`Unknown item type: ${item.type}`);
assertGroupsMatchType(item);
const version = item.v ?? ITEM_SCHEMA_VERSION;
const plaintext = utf8Encode(JSON.stringify({ ...item, v: version }));
const { iv, ciphertext } = await aesGcmEncrypt(userKey, plaintext, itemAad(namespace, item.id, version));
return {
id: item.id,
type: ITEM_TYPE[item.type],
ciphertext: toBase64(ciphertext),
iv: toBase64(iv),
};
}
/**
* Decrypt a stored envelope.
*
* Throws when the key is wrong, the ciphertext was altered, or the row's id
* does not match the one bound at encryption time.
*/
export async function decryptItem(
userKey: Uint8Array,
row: Envelope,
namespace: Namespace = DEFAULT_NAMESPACE,
): Promise<Item> {
const version = row.v ?? ITEM_SCHEMA_VERSION;
let plaintext: Uint8Array;
try {
plaintext = await aesGcmDecrypt(
userKey,
fromBase64(row.iv),
fromBase64(row.ciphertext),
itemAad(namespace, row.id, version),
);
} catch {
throw new Error(`Could not decrypt item ${row.id}`);
}
const item = JSON.parse(utf8Decode(plaintext)) as Item;
if (item.id !== row.id) {
// Belt and braces: the AAD already makes this unreachable.
throw new Error(`Item id mismatch for ${row.id}`);
}
return item;
}
export interface DecryptResult {
items: Item[];
failed: Array<{ id: string; error: string }>;
}
/**
* Decrypt a page of rows, keeping going when one fails.
*
* A single corrupt row must not hide the rest of someone's vault, so failures
* are collected and returned rather than thrown.
*/
export async function decryptItems(
userKey: Uint8Array,
rows: Envelope[],
namespace: Namespace = DEFAULT_NAMESPACE,
): Promise<DecryptResult> {
const items: Item[] = [];
const failed: Array<{ id: string; error: string }> = [];
for (const row of rows) {
try {
items.push(await decryptItem(userKey, row, namespace));
} catch (err) {
failed.push({ id: row.id, error: (err as Error).message });
}
}
return { items, failed };
}
/** The field paths that hold a secret, per type. Used for masking and reveal. */
export const SECRET_FIELDS: Readonly<Record<ItemTypeName, readonly string[]>> = Object.freeze({
login: ["login.password", "login.totp"],
card: ["card.number", "card.code"],
identity: ["identity.ssn", "identity.passportNumber", "identity.licenseNumber"],
note: [],
key: ["key.privateKey", "key.passphrase", "key.value"],
account: ["account.accessToken", "account.refreshToken"],
});
/**
* A copy of an item with every secret replaced by a mask.
*
* Used by every display path, including `--json`: a pipeline is not an
* authorization, and an item that prints its password when redirected to a file
* is an item that leaks into shell history and CI logs.
*/
export function maskItem(item: Item, mask = "••••••••"): Item {
const copy = structuredClone(item) as Item;
for (const path of SECRET_FIELDS[item.type] ?? []) {
const [group, field] = path.split(".") as [string, string];
const holder = (copy as Record<string, unknown>)[group] as Record<string, unknown> | undefined;
if (holder && typeof holder[field] === "string" && holder[field] !== "") holder[field] = mask;
}
if (Array.isArray(copy.fields)) {
copy.fields = copy.fields.map((f) => (f.type === "hidden" && f.value ? { ...f, value: mask } : f));
}
if (Array.isArray(copy.history) && copy.history.length > 0) {
copy.history = copy.history.map((h) => ({ ...h, password: mask }));
}
if (Array.isArray(copy.attachments)) {
copy.attachments = copy.attachments.map(({ key, ...rest }) => (key ? { ...rest, key: mask } : rest));
}
return copy;
}
/** Read one field by dotted path, for a deliberate single-value reveal. */
export function readField(item: Item, path: string): string | undefined {
const parts = path.split(".");
let cursor: unknown = item;
for (const part of parts) {
if (typeof cursor !== "object" || cursor === null) return undefined;
cursor = (cursor as Record<string, unknown>)[part];
}
return typeof cursor === "string" ? cursor : cursor === undefined ? undefined : JSON.stringify(cursor);
}