Merge pull request 'feat(comments): attribute MCP agent comments as AI (unspoofable provenance)' (#143) from feat/mcp-comments-ai-attribution into develop

Reviewed-on: #143
This commit was merged in pull request #143.
This commit is contained in:
2026-06-24 02:05:06 +03:00
27 changed files with 965 additions and 361 deletions

View File

@@ -1,3 +1,11 @@
/**
* Provenance actor for a write: who the action is attributed to. Derived only
* from the SIGNED token claim (never a request body), so 'agent' is unspoofable.
* Single source of truth so a typo like 'agnet' can't slip through as a bare
* string (#143 review). Distinct from `ActorType` (auth principal kind).
*/
export type ProvenanceSource = 'user' | 'agent';
export enum JwtType {
ACCESS = 'access',
COLLAB = 'collab',
@@ -19,8 +27,10 @@ export type JwtPayload = {
// mints a provenance access token so REST writes (create/rename/move page,
// comment create/resolve) record a non-spoofable 'agent' marker (§6.5 / §15
// C3 / §14 N2).
actor?: 'user' | 'agent';
aiChatId?: string;
actor?: ProvenanceSource;
// Nullable: an external MCP agent has no internal ai_chats row, so it carries
// an 'agent' actor with a null aiChatId.
aiChatId?: string | null;
};
export type JwtCollabPayload = {
@@ -30,8 +40,10 @@ export type JwtCollabPayload = {
// Optional agent-edit provenance, signed into the collab token. Absent for
// the human collab path (treated as 'user'); set only when the internal agent
// mints a provenance collab token (§6.6 / §15 C2).
actor?: 'user' | 'agent';
aiChatId?: string;
actor?: ProvenanceSource;
// Nullable: an external MCP agent has no internal ai_chats row, so it carries
// an 'agent' actor with a null aiChatId.
aiChatId?: string | null;
};
export type JwtExchangePayload = {

View File

@@ -34,7 +34,9 @@ export class TokenService {
// token carries no actor/aiChatId and is treated as 'user' downstream. The
// internal agent passes { actor:'agent', aiChatId } so REST writes record a
// non-spoofable 'agent' marker off the signed claim (§6.5 / §15 C3 / §14 N2).
provenance?: { actor: 'agent'; aiChatId: string },
// aiChatId is nullable: an external MCP agent has no internal ai_chats row,
// so it stamps 'agent' with a null aiChatId.
provenance?: { actor: 'agent'; aiChatId: string | null },
): Promise<string> {
if (isUserDisabled(user)) {
throw new ForbiddenException();
@@ -58,7 +60,8 @@ export class TokenService {
workspaceId: string,
// Optional agent-edit provenance. When omitted (the human collab path), the
// token carries no actor/aiChatId and is treated as 'user' downstream.
provenance?: { actor: 'agent'; aiChatId: string },
// aiChatId is nullable for an external agent with no internal ai_chats row.
provenance?: { actor: 'agent'; aiChatId: string | null },
): Promise<string> {
if (isUserDisabled(user)) {
throw new ForbiddenException();

View File

@@ -0,0 +1,122 @@
import { UnauthorizedException } from '@nestjs/common';
import { JwtStrategy } from './jwt.strategy';
import { JwtType } from '../dto/jwt-payload';
/**
* Provenance derivation in JwtStrategy.validate (jwt.strategy.ts).
*
* The strategy must derive the agent-edit provenance from the SIGNED server-side
* identity, never from a client-controlled field. The security invariant under
* test: a user flagged is_agent stamps 'agent'; an ordinary user resolves to
* 'user'; and an `actor` claim in the token CANNOT escalate a non-agent user
* past the existing internal-AI-chat claim semantics (anti-spoof — a plain user
* cannot obtain created_source='agent').
*
* The strategy is constructed directly with stub deps. The PassportStrategy base
* only needs a secret at construction time; validate() is exercised on its own.
*/
describe('JwtStrategy — provenance derivation', () => {
function makeStrategy(user: any) {
const userRepo: any = { findById: jest.fn(async () => user) };
const workspaceRepo: any = { findById: jest.fn(async () => ({ id: 'ws-1' })) };
const userSessionRepo: any = { findActiveById: jest.fn() };
const sessionActivityService: any = { trackActivity: jest.fn() };
const environmentService: any = { getAppSecret: () => 'test-secret' };
const moduleRef: any = {};
const strategy = new JwtStrategy(
userRepo,
workspaceRepo,
userSessionRepo,
sessionActivityService,
environmentService,
moduleRef,
);
return { strategy, userRepo };
}
// A bare request whose `raw` collects the provenance the strategy stamps.
const makeReq = () => ({ raw: {} as Record<string, any> });
const accessPayload = (over?: Record<string, any>) => ({
sub: 'user-1',
email: 'u@test.local',
workspaceId: 'ws-1',
type: JwtType.ACCESS,
...over,
});
it("stamps actor='agent' for an is_agent user (derived from the signed identity)", async () => {
const { strategy, userRepo } = makeStrategy({
id: 'user-1',
isAgent: true,
deactivatedAt: null,
deletedAt: null,
});
const req = makeReq();
await strategy.validate(req, accessPayload() as any);
expect(req.raw.actor).toBe('agent');
// External MCP agent: no internal ai_chats row → null.
expect(req.raw.aiChatId).toBeNull();
// Wiring guard (#143): the seam MUST opt into the isAgent flag, otherwise
// findById omits it (it is not in baseFields) and provenance silently
// degrades to 'user'.
expect(userRepo.findById).toHaveBeenCalledWith(
'user-1',
'ws-1',
expect.objectContaining({ includeIsAgent: true }),
);
});
it("stamps actor='user' for an ordinary user", async () => {
const { strategy } = makeStrategy({
id: 'user-1',
isAgent: false,
deactivatedAt: null,
deletedAt: null,
});
const req = makeReq();
await strategy.validate(req, accessPayload() as any);
expect(req.raw.actor).toBe('user');
expect(req.raw.aiChatId).toBeNull();
});
it("honors a SIGNED actor='agent' claim on a non-agent user's token (the internal AI-chat path)", async () => {
// A non-agent user (the plain no-claim → 'user' case is covered above). A
// token that DOES carry actor='agent' resolves to 'agent' — BY DESIGN: that
// claim can only exist on a SERVER-MINTED provenance token (the internal AI
// chat), never on a plain login token, because the token is signed with the
// app secret. The guarantee is that a client cannot FORGE this signed claim,
// not that the strategy ignores it. (A plain user still cannot obtain
// 'agent' — they have no way to get such a token.)
const { strategy } = makeStrategy({
id: 'user-1',
isAgent: false,
deactivatedAt: null,
deletedAt: null,
});
const req2 = makeReq();
await strategy.validate(req2, accessPayload({ actor: 'agent', aiChatId: 'chat-1' }) as any);
expect(req2.raw.actor).toBe('agent');
expect(req2.raw.aiChatId).toBe('chat-1');
});
it('rejects a disabled is_agent user (Unauthorized) before stamping provenance', async () => {
const { strategy } = makeStrategy({
id: 'user-1',
isAgent: true,
deactivatedAt: new Date('2026-01-01'),
deletedAt: null,
});
const req = makeReq();
await expect(strategy.validate(req, accessPayload() as any)).rejects.toThrow(
UnauthorizedException,
);
expect(req.raw.actor).toBeUndefined();
});
});

View File

@@ -10,6 +10,7 @@ import { SessionActivityService } from '../../session/session-activity.service';
import { FastifyRequest } from 'fastify';
import { extractBearerTokenFromHeader, isUserDisabled } from '../../../common/helpers';
import { ModuleRef } from '@nestjs/core';
import { resolveProvenance } from '../../../common/decorators/auth-provenance.decorator';
@Injectable()
export class JwtStrategy extends PassportStrategy(Strategy, 'jwt') {
@@ -55,7 +56,9 @@ export class JwtStrategy extends PassportStrategy(Strategy, 'jwt') {
if (!workspace) {
throw new UnauthorizedException();
}
const user = await this.userRepo.findById(payload.sub, payload.workspaceId);
const user = await this.userRepo.findById(payload.sub, payload.workspaceId, {
includeIsAgent: true,
});
if (!user || isUserDisabled(user)) {
throw new UnauthorizedException();
@@ -71,14 +74,15 @@ export class JwtStrategy extends PassportStrategy(Strategy, 'jwt') {
this.sessionActivityService.trackActivity(sessionId, payload.sub, payload.workspaceId);
}
// Propagate the signed agent-edit provenance claim onto the request so REST
// services/controllers can set the 'agent' marker off it. A normal user
// token carries no actor claim and resolves to 'user' (unchanged behaviour);
// only the internal agent's minted token sets actor='agent' + aiChatId. This
// is read server-side from the SIGNED token, never from a client body field,
// so a normal user cannot fake an 'agent' badge.
req.raw.actor = (payload as JwtPayload).actor ?? 'user';
req.raw.aiChatId = (payload as JwtPayload).aiChatId ?? null;
// Propagate the agent-edit provenance onto the request so REST
// services/controllers can set the 'agent' marker off it. Derived from the
// SIGNED server-side identity via the shared resolver (also used by the
// collab seam, so the two never drift), never from a client body field — so
// an is_agent service account stamps every REST write made with an access
// token, and a normal user cannot fake an 'agent' badge.
const provenance = resolveProvenance(user, payload as JwtPayload);
req.raw.actor = provenance.actor;
req.raw.aiChatId = provenance.aiChatId;
return { user, workspace };
}

View File

@@ -147,6 +147,24 @@ describe('CommentService — behavior', () => {
expect(insertArg.creatorId).toBe('user-1');
});
it('stamps createdSource:"agent" with a null aiChatId (external MCP agent) without breaking insert', async () => {
const { service, commentRepo } = makeService();
// An external MCP agent is flagged is_agent server-side but has no
// internal ai_chats row, so provenance carries actor='agent' + a null
// aiChatId. The insert must still record the agent marker.
await service.create(
{ page: page(), workspaceId: 'ws-1', user: user() },
{ content: JSON.stringify(docMentioning()) } as any,
{ actor: 'agent', aiChatId: null },
);
const insertArg = commentRepo.insertComment.mock.calls[0][0];
expect(insertArg.createdSource).toBe('agent');
expect(insertArg.aiChatId).toBeNull();
expect(insertArg.creatorId).toBe('user-1');
});
it('leaves source default (no agent stamp) for a normal user', async () => {
const { service, commentRepo } = makeService();

View File

@@ -22,7 +22,10 @@ import {
ICommentResolvedNotificationJob,
} from '../../integrations/queue/constants/queue.interface';
import { WsService } from '../../ws/ws.service';
import { AuthProvenanceData } from '../../common/decorators/auth-provenance.decorator';
import {
AuthProvenanceData,
agentSourceFields,
} from '../../common/decorators/auth-provenance.decorator';
@Injectable()
export class CommentService {
@@ -60,7 +63,6 @@ export class CommentService {
) {
const { page, workspaceId, user } = opts;
const commentContent = JSON.parse(createCommentDto.content);
const isAgent = provenance?.actor === 'agent';
if (createCommentDto.parentCommentId) {
const parentComment = await this.commentRepo.findById(
@@ -87,9 +89,7 @@ export class CommentService {
spaceId: page.spaceId,
// Agent-edit provenance: the user stays creatorId; this only annotates the
// source. Normal user requests leave the column default ('user').
...(isAgent
? { createdSource: 'agent', aiChatId: provenance.aiChatId }
: {}),
...agentSourceFields(provenance, 'createdSource', 'aiChatId'),
});
if (createCommentDto.yjsSelection) {

View File

@@ -147,4 +147,246 @@ describe('PageService', () => {
expect(pageRepo.updatePage).toHaveBeenCalledTimes(1);
});
});
describe('agent provenance stamping (#143)', () => {
// Provenance handed to the four write sites. The agent case must surface the
// signed source marker + chat id on the persisted payload; the user case must
// leave both keys absent so the column keeps its INSERT default / existing
// UPDATE value (agentSourceFields returns {} for a non-agent).
const AGENT = { actor: 'agent', aiChatId: 'chat-7' } as any;
const USER = { actor: 'user', aiChatId: null } as any;
// A general-queue stub whose `.add(...)` returns a `{ catch }` thenable —
// the service does `generalQueue.add(...).catch(...)` and never awaits it.
const makeGeneralQueue = () =>
({ add: jest.fn().mockReturnValue({ catch: jest.fn() }) }) as any;
// Build a PageService where only the deps a given site touches are real
// stubs; everything else stays a bare object. db is supplied per-test.
const makeSvc = (overrides: {
pageRepo?: any;
generalQueue?: any;
db?: any;
}) =>
new PageService(
(overrides.pageRepo ?? {}) as any, // pageRepo
{} as any, // pagePermissionRepo
{} as any, // attachmentRepo
(overrides.db ?? {}) as any, // db
{} as any, // storageService
{} as any, // attachmentQueue
{} as any, // aiQueue
(overrides.generalQueue ?? makeGeneralQueue()) as any, // generalQueue
{} as any, // eventEmitter
{} as any, // collaborationGateway
{} as any, // watcherService
{} as any, // transclusionService
);
describe('create() → insertPage', () => {
const run = async (provenance: any) => {
const pageRepo = {
insertPage: jest.fn().mockResolvedValue({ id: 'p1' }),
};
const svc = makeSvc({ pageRepo, generalQueue: makeGeneralQueue() });
// nextPagePosition runs a real db query; stub it out.
jest.spyOn(svc, 'nextPagePosition').mockResolvedValue('a0' as any);
// No content/format → the prosemirror parse branch is skipped. No
// parentPageId → no parent lookup.
await svc.create(
'u1',
'w1',
{ title: 't', spaceId: 's1' } as any,
provenance,
);
return pageRepo.insertPage.mock.calls[0][0];
};
it('stamps lastUpdatedSource/lastUpdatedAiChatId for an agent', async () => {
const payload = await run(AGENT);
expect(payload).toEqual(
expect.objectContaining({
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: 'chat-7',
}),
);
});
it('omits the source columns for a normal user', async () => {
const payload = await run(USER);
expect(payload).not.toHaveProperty('lastUpdatedSource');
expect(payload).not.toHaveProperty('lastUpdatedAiChatId');
});
});
describe('update() → updatePage', () => {
const run = async (provenance: any) => {
const pageRepo = {
updatePage: jest.fn().mockResolvedValue(undefined),
findById: jest.fn().mockResolvedValue({ id: 'p1' }),
};
const svc = makeSvc({ pageRepo, generalQueue: makeGeneralQueue() });
const page = {
id: 'p1',
contributorIds: [],
spaceId: 's1',
workspaceId: 'w1',
slugId: 'sl1',
title: 't',
parentPageId: null,
} as any;
// dto carries no content/operation/format → updatePageContent skipped.
await svc.update(page, {} as any, { id: 'u1' } as any, provenance);
return pageRepo.updatePage.mock.calls[0][0];
};
it('stamps lastUpdatedSource/lastUpdatedAiChatId for an agent', async () => {
const payload = await run(AGENT);
expect(payload).toEqual(
expect.objectContaining({
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: 'chat-7',
}),
);
});
it('omits the source columns for a normal user', async () => {
const payload = await run(USER);
expect(payload).not.toHaveProperty('lastUpdatedSource');
expect(payload).not.toHaveProperty('lastUpdatedAiChatId');
});
});
describe('movePage() → updatePage', () => {
const VALID_POSITION = 'a0';
const run = async (provenance: any) => {
const pageRepo = {
findById: jest.fn().mockResolvedValue({
id: 'dest-parent',
deletedAt: null,
spaceId: 'space-1',
}),
updatePage: jest.fn().mockResolvedValue({ numUpdatedRows: 1n }),
};
const svc = makeSvc({
pageRepo,
db: {} as any,
});
// Legitimate move: destination ancestors do NOT include the moved page.
jest
.spyOn(svc, 'getPageBreadCrumbs')
.mockResolvedValue([{ id: 'dest-parent' }, { id: 'root' }] as any);
// eventEmitter is a bare {} stub; movePage emits PAGE_MOVED, so give it
// an emit. Re-wire via the private field to avoid threading it through.
(svc as any).eventEmitter = { emit: jest.fn() };
const movedPage = {
id: 'page-1',
parentPageId: 'old-parent',
spaceId: 'space-1',
workspaceId: 'ws-1',
slugId: 'slug-1',
title: 'Page 1',
icon: null,
} as any;
const dto = {
pageId: 'page-1',
position: VALID_POSITION,
parentPageId: 'dest-parent',
} as any;
await svc.movePage(dto, movedPage, provenance);
return pageRepo.updatePage.mock.calls[0][0];
};
it('stamps lastUpdatedSource/lastUpdatedAiChatId for an agent', async () => {
const payload = await run(AGENT);
expect(payload).toEqual(
expect.objectContaining({
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: 'chat-7',
}),
);
});
it('omits the source columns for a normal user', async () => {
const payload = await run(USER);
expect(payload).not.toHaveProperty('lastUpdatedSource');
expect(payload).not.toHaveProperty('lastUpdatedAiChatId');
});
});
describe('movePageToSpace() → root-page updatePage', () => {
// movePageToSpace runs its writes inside executeTx(this.db, cb), which
// calls this.db.transaction().execute(fn => fn(trx)). A permissive
// chainable Proxy stands in for the Kysely trx so arbitrary chains resolve.
const makeChain = () => {
const c: any = new Proxy(function () {}, {
get: (_t, p) =>
p === 'then'
? undefined
: p === 'execute' || p === 'executeTakeFirst'
? () => Promise.resolve([])
: () => c,
});
return c;
};
const run = async (provenance: any) => {
const trxStub = makeChain();
const db = {
transaction: () => ({ execute: (fn: any) => fn(trxStub) }),
} as any;
const rootPage = {
id: 'root',
spaceId: 'src-space',
parentPageId: null,
workspaceId: 'ws-1',
} as any;
const pageRepo = {
getPageAndDescendants: jest.fn().mockResolvedValue([rootPage]),
updatePage: jest.fn().mockResolvedValue(undefined),
updatePages: jest.fn().mockResolvedValue(undefined),
};
const svc = makeSvc({ pageRepo, db });
// The single-accessible-page path still runs the bulk side-effect writes
// (attachments/watchers/ai-queue) AFTER the root updatePage we assert on;
// stub them so the transaction completes without throwing.
(svc as any).attachmentRepo = {
updateAttachmentsByPageId: jest.fn().mockResolvedValue(undefined),
};
(svc as any).watcherService = {
movePageWatchersToSpace: jest.fn().mockResolvedValue(undefined),
};
(svc as any).aiQueue = { add: jest.fn().mockResolvedValue(undefined) };
// Single accessible page (the root) → pagesToOrphan is empty, so the
// root updatePage is the first/only provenance-carrying updatePage call.
// filterAccessibleTreePages is private; spy via an `any` cast.
jest
.spyOn(svc as any, 'filterAccessibleTreePages')
.mockResolvedValue([rootPage] as any);
jest.spyOn(svc, 'nextPagePosition').mockResolvedValue('a0' as any);
await svc.movePageToSpace(rootPage, 'dst-space', 'u1', provenance);
return pageRepo.updatePage.mock.calls[0][0];
};
it('stamps the moved root with the agent source + chat id', async () => {
const payload = await run(AGENT);
expect(payload).toEqual(
expect.objectContaining({
spaceId: 'dst-space',
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: 'chat-7',
}),
);
});
it('omits the source columns on the moved root for a normal user', async () => {
const payload = await run(USER);
expect(payload).toEqual(
expect.objectContaining({ spaceId: 'dst-space' }),
);
expect(payload).not.toHaveProperty('lastUpdatedSource');
expect(payload).not.toHaveProperty('lastUpdatedAiChatId');
});
});
});
});

View File

@@ -57,7 +57,10 @@ import { WatcherService } from '../../watcher/watcher.service';
import { sql } from 'kysely';
import { TransclusionService } from '../transclusion/transclusion.service';
import { remapPageEmbedSourceId } from '../transclusion/utils/transclusion-prosemirror.util';
import { AuthProvenanceData } from '../../../common/decorators/auth-provenance.decorator';
import {
AuthProvenanceData,
agentSourceFields,
} from '../../../common/decorators/auth-provenance.decorator';
@Injectable()
export class PageService {
@@ -135,8 +138,6 @@ export class PageService {
ydoc = createYdocFromJson(prosemirrorJson);
}
const isAgent = provenance?.actor === 'agent';
const page = await this.pageRepo.insertPage({
slugId: generateSlugId(),
title: createPageDto.title,
@@ -153,12 +154,7 @@ export class PageService {
// Agent-edit provenance. The human stays the responsible author
// (creatorId/lastUpdatedById); these only annotate the source. A normal
// user request leaves the column default ('user').
...(isAgent
? {
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: provenance.aiChatId,
}
: {}),
...agentSourceFields(provenance, 'lastUpdatedSource', 'lastUpdatedAiChatId'),
content,
textContent,
ydoc,
@@ -231,8 +227,6 @@ export class PageService {
contributors.add(user.id);
const contributorIds = Array.from(contributors);
const isAgent = provenance?.actor === 'agent';
// Detect a real title/icon change so the WS tree listener can broadcast an
// `updateOne` to the space (rename / icon swap) WITHOUT re-broadcasting on a
// content-only save. Only treat a field as changed when the DTO actually
@@ -250,13 +244,9 @@ export class PageService {
icon: updatePageDto.icon,
lastUpdatedById: user.id,
// Agent-edit provenance: annotate the source without changing the
// responsible author. A normal user request leaves the column default.
...(isAgent
? {
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: provenance.aiChatId,
}
: {}),
// responsible author. A normal user request leaves the existing source
// value unchanged.
...agentSourceFields(provenance, 'lastUpdatedSource', 'lastUpdatedAiChatId'),
updatedAt: new Date(),
contributorIds: contributorIds,
},
@@ -443,7 +433,6 @@ export class PageService {
provenance?: AuthProvenanceData,
) {
let childPageIds: string[] = [];
const isAgent = provenance?.actor === 'agent';
const allPages = await this.pageRepo.getPageAndDescendants(rootPage.id, {
includeContent: false,
@@ -490,12 +479,7 @@ export class PageService {
// Agent-edit provenance on the moved root page. Child pages are bulk
// re-parented to the new space (no content change), so the marker is
// stamped on the root the agent acted on. Normal user: no change.
...(isAgent
? {
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: provenance.aiChatId,
}
: {}),
...agentSourceFields(provenance, 'lastUpdatedSource', 'lastUpdatedAiChatId'),
},
rootPage.id,
trx,
@@ -949,20 +933,13 @@ export class PageService {
}
}
const isAgent = provenance?.actor === 'agent';
const updateResult = await this.pageRepo.updatePage(
{
position: dto.position,
parentPageId: parentPageId,
// Agent-edit provenance: annotate the source on an agent move. A normal
// user request leaves the column default ('user').
...(isAgent
? {
lastUpdatedSource: 'agent',
lastUpdatedAiChatId: provenance.aiChatId,
}
: {}),
// user request leaves the existing source value unchanged.
...agentSourceFields(provenance, 'lastUpdatedSource', 'lastUpdatedAiChatId'),
},
dto.pageId,
);