diff --git a/agents/rocko/work/queue-219/BUILD.md b/agents/rocko/work/queue-219/BUILD.md new file mode 100644 index 000000000..aecd29ba3 --- /dev/null +++ b/agents/rocko/work/queue-219/BUILD.md @@ -0,0 +1,181 @@ +# Row 219 / C2-09a: one thread per creative instance + +Assignment sage-2793, base aca16404c; branch rocko/row219, own worktree +/mnt/storage/scratch/rocko-r219. Claim waits for Sage's gap. No Node, tests, +HTTPS git or push while Gander's row253 gate holds the slot. + +Read both C2-09a and C2-09b briefs at aca16404c and 51cc0445c: the two +sections are byte-identical. The newer brief changes C2-06's reply limits +under Q222, not C2-09a. This row introduces no M5 limit, writer or setting. +Row265 rooms chains on this row and stays separate. + +## Implementation plan + +Broker: replace assistant-only #threaded with THREAD_DEFINITIONS membership. +thread.create gains optional instance, validates it as a declared thread +instance, and binds both the thread and default session to it. Omitted +instance retains single-instance fallback and ambiguous/missing refusals. +User grants must still cover the selected instance and project before +project lookup; an explicit selector cannot bypass FED R3. Existing tester +ownership and per-thread/per-session holder checks remain. + +Launcher: import shared THREAD_DEFINITIONS and replace its two assistant +checks (check and ending), preserving launch admission, quotas, session +isolation and stop policy. No runtime.mjs, bundle or library adapter change; +Della row224 was notified of these narrow launcher hunks. + +Console: actual shipped new-thread UI is in public/chat.js, with transport +mapping in chat.mjs and HTTP validation in chat-http.mjs, not public/bus.js. +These three files need Sage's scope addition before their edits. Proposed +metadata read GET /api/chat/instances in serve.mjs uses the existing +bus.instances read, filters using shared THREAD_DEFINITIONS, and returns +only {instance,name}, with vars['persona.name'] or instance as the label. +It adds no channel operation or bus verb; send still traverses createChat +and the existing channel dispatch. The new select sends optional instance +through chat-http validation and chat.mjs. It is disabled when metadata +cannot be read, so the UI never silently chooses between agents. + +## Acceptance evidence to build (runs await CLEAR) + +Bus: explicit creative selection binds the requested thread and default +session; omitted selector preserves assistant compatibility, missing and +ambiguous cases. Unknown/non-thread/malformed selectors refuse without +new threads; user grants cover only the selected instance, with prior +non-disclosure order intact. Real creative role definition messaging still +holds. A creative thread claims/receives independently of its peer. + +CLI: real fixture launcher starts a creative thread on pending input and +keeps peer thread/session stores separate. Direct non-thread launch of a +creative instance refuses thread-required. Existing assistant tests remain. + +WebUI: metadata lists thread definitions only with persona/name labels; +selection survives refresh and reaches thread.create exactly. Malformed +instance fails before a write, old omitted instance still works. Browser +asserts two creative options, inert hostile labels and read-only disablement. +Keep all identity/origin/content/SSO fences. Named mutants cover each new +acceptance guard; no runs until Sage's announced gap/CLEAR. + +Before review: grep source readers/hash pins, update only pins required by +owned changes, rebase onto row253 landing under a new branch, then rerun +touched files on the new tree. No schema, role-policy, live or canonical +product-path change. Handoff directly to Dewey, Mos and Sage when ready. + +## Scope ruling and picker design before UI edits + +Sage-2807 adds public/chat.js, chat.mjs and chat-http.mjs because the shipped +form is chat.js:55/235, its forwarding is chat.mjs:23, and POST validation is +chat-http.mjs:80-82. bus.js stays byte-identical. The new route sits after +existing auth/session checks and uses startServer's own bus, so the tester +socket's separately supplied testerBus is used too (Dewey U1). No new bus or +channel operation. The route projects only instance/name and keeps the +THREAD_DEFINITIONS filter server-side, rather than making the browser read +all /api/bus/instances vars (Dewey route note). + +Dewey-0849 U2-U5: multiple choices start with a required placeholder; exactly +one starts chosen and is sent explicitly; none disables Create and says no +thread instances are configured. Failed metadata reads disable Create with +"Couldn't read the thread instances; reload to try again" and recover on a +good read. Labels show persona name plus instance key when different. Picks +live in form(key); a removed pick is cleared with a status and never moved +to another option, even if only one remains. Repaints retain selection. +U6 is deferred: thread-list/header still show their existing instance keys. +The shared sign-out/redirect handling and strict HTTP write key validation +remain in force. Add unauthenticated metadata and selected-instance user +grant refusal tests as Sage requires. + +## Existing fixture adaptations (unrun) + +C2-08 creative.test's setup previously bound creative roles without threads +and sent human input with message.send. This row makes those roles threaded: +setup now selects assistant explicitly, creates each creative instance's +thread and binds its claim there; the human uses thread.send. Existing +M6/messaging refusal and count assertions stay in place. The user-principal +fixture gains optional extra threadRoles for the selected-grant regression. + +The browser fixture now answers /api/chat/instances with its one assistant. +Its existing create receipt assertion gains instance:'assistant', since the +new picker explicitly sends its single preselected choice. No existing +assertion is dropped. New tests cover multiple/empty/failed metadata, +duplicate hostile labels, selected-instance POST, refresh and disappearing +choices. No Node or test has been run under Gander's hold. + +The existing launcher.test C2-08 persona-mismatch regression cannot keep +requesting a standalone creative launch: C2-09a now correctly refuses it +thread-required before bundle construction. Its trigger is therefore the +thread pass, with an injected host threadsState snapshot (the persona +mismatch occurs before any bind). The exact mismatch reason, no persona +content leak, one refusal and no running session assertions remain. The +old per-run `['session']` directory assertion changes to `[]` and gains an +assertion for the thread-owned session directory, matching row121's actual +thread-store layout. A new e.thread assertion pins the path. This is a +fixture/expected-layout adaptation, not a product exception for that test. + +## Old-test approval: mos-1316 + +Mos approved the three adaptations in OLD-TEST-CHANGES.md (sha256 +9a1e51e6b43040593a3dd517c4a0c71f5ac42b6b57d66d474cd0517f06e97ed6) +and old-persona-test.diff (sha256 +6aaca9820b692ec659058e9b044f7f60872cdecb16fe3ac1321c2c0ed530701d). +The source commit must name mos-1316 and carry OLD-TEST-CHANGES.md. +Dewey checks each changed line against that record. + +Keep the new thread-launch test's fromSession(chief.cap, 'bean') refusal +(thread-required), and the threads test's direct human message.send refusal +for each creative definition (thread-required-use-thread-send). The former +is sandbox-dependent: its gate record must show it ran without a skip; +otherwise report to Mos rather than claiming the pin. T06 restores the old +assistant-only launcher check; T22 restores the ending check separately. +The persona mismatch fixture wraps only threadsState; all other host ops +remain real. Under mos-1272, a first-run red in launcher.test.mjs, +creative.test.mjs or chat-browser.test.mjs goes to Mos before any fix. +No run has occurred as of this approval record. + +### mos-1317: always-running standalone admission pin + +Added a separate new test next to the assistant standalone refusal: +`C2-09a: fromSession refuses a creative instance outside a thread without a sandbox`. +It has no sandbox skip. The real fixture host binds/claims the chief, and +fromSession must refuse Bean with thread-required before bundle/sandbox +construction. The existing sandbox-dependent pin remains unchanged. +T23 restores both launcher assistant-only comparisons and targets this new +exact refusal assertion. T06 and T22 remain as separate launch/ending checks: +23 mutants total. No tests run; slot scope amendment sent to Sage. +Keep the Mos-approved threadsState wrap as the only persona-test host override. + +## Inspection candidate under sage-3113 (2026-10-23) + +Explicitly UNVERIFIED. Sage authorized a local inspection commit before +runs; no claim of review readiness or green tests. Base remains aca16404c, +branch rocko/row219. No rebase has been performed. No Node, browser, test, +mutant or suite has run; the earlier gate holds and subsequent row219 hold +were honored. Static Python exact-once preparation named 23 mutant sites; +git diff --check passes. The mutant execution runner is not yet written. + +14 source/test paths: six product files and eight whole test files. The +candidate manifest excludes this packet. The new no-sandbox fromSession +Bean test and T23 record mos-1317; old test adaptations stay governed by +mos-1316 and OLD-TEST-CHANGES.md. First-run reds in launcher, creative and +chat-browser must go to Mos before any fix. Run evidence must prove the +sandbox-dependent pin ran, or report its skip to Mos. + +Remaining: Sage-directed rebase to current landing base/new branch; slot +CLEAR and claim; eight whole test files and 23 mutants with process-group +cleanup; path/hash-reader sweep, acceptance evidence and final review packet. +No product behavior is verified. Push waits for Sage's named gap; reviewers +and dependent rows may inspect this commit but must not treat it as approved. + +## Rebase under sage-3127 / exact resolution sage-3128 + +Rebased the unverified c4b4fab98 inspection candidate from aca16404c onto +062646487451993302c768c6812b827c834e952b. The only conflict was broker.mjs's +thread.create user guard: row206 added verb-sensitive #granted(s, verb, +project, instance). Sage-3128 supplied the exact resolution: retain the +selected-instance validation and call #granted(s, 'thread.create', +identifier(a.project), instance). No guessed conflict resolution. The merged +broker has no remaining three-argument #granted(s, ...) call (static balanced +argument scan). T05 now matches that four-argument call and substitutes null +only for instance; all 23 mutant old anchors, including T23's second edit, +remain exact-once. Rebuild manifest/base/patch on this base. Other auto-merges +were user-principal.test, launcher and thread-launch.test. No tests, Node, +mutants or reader audit run: CLEAR still required after 574, covering every +touched test file whole. This branch remains UNVERIFIED. diff --git a/agents/rocko/work/queue-219/OLD-TEST-CHANGES.md b/agents/rocko/work/queue-219/OLD-TEST-CHANGES.md new file mode 100644 index 000000000..ecc309cd6 --- /dev/null +++ b/agents/rocko/work/queue-219/OLD-TEST-CHANGES.md @@ -0,0 +1,58 @@ +# Old-test changes awaiting Mos under mos-1272 (sage-2822) + +No test has run. Row219's product patch is uncommitted on rocko/row219, +aca16404c, /mnt/storage/scratch/rocko-r219. + +## Persona-mismatch regression: launcher.test.mjs + +Name: `row 213 a creative instance's persona is read from /personas: +off its digest the launch fails with the mismatch as its reason`. + +Old trigger: bind/claim a chief, then +`assert.deepEqual(await launcher.fromSession(chief.cap, "bean"), { ok: false, error: "launch-failed" });`. + +New trigger: fixture-wrap only `host.op({op:'threadsState'})` to report one +pending Bean thread, then `await launcher.tick()`. No fake grant, bind, +writer or persona reader. The same actual persona bundle check fails before +any thread run binds. All other host operations use the real fixture host. + +Why: C2-08 declares companion and the shared THREAD_DEFINITIONS. Row219 now +uses that list in check(). Bean's companion definition is therefore +threaded, and fromSession's standalone request refuses thread-required +before start()/buildBundle can read the persona. Keeping that trigger would +stop testing the persona check. No product exemption is added. + +Preserved assertions: exactly one refusal; event/code = refused/launch-failed; +exact persona-mismatch reason including the fixture persona path; no +PERSONA-MARKER-d2 in logs; no running sessions; exactly one failed run dir. +The direct fromSession return assertion is replaced by a thread-id assertion +on the same refusal record. Both prove the refusal at the appropriate entry. + +Old layout assertion: +`assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), ["session"]);` + +New layout assertions: +`assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), []);` +`assert.equal(existsSync(join(f.dataRoot, "threads", "acme", thread, "session")), true);` + +This matches row121: a thread uses its thread-owned store, never a per-run +session directory. The failed persona build still writes no bundle files. +Exact diff: old-persona-test.diff. + +## Other existing-test adaptations disclosed with it + +- bus creative.test setup now explicitly chooses assistant, creates Bean and + Quill threads, and binds those runs with their own thread ids. The existing + human `message.send {to:who}` input becomes `thread.send {thread:threads[who]}`. + All original M6 decision, messaging refusal and message-count assertions + remain unchanged. This lets those same restrictions run on the now-threaded + creative roles; direct human messages are correctly thread-required. +- webui chat-browser.test fixture adds /api/chat/instances metadata. Its old + expected create body `{project:P,name:'named thread'}` becomes + `{project:P,name:'named thread',instance:'assistant'}` because the one-choice + picker explicitly sends its preselected instance (Dewey U2). Every other + existing assertion remains. The additional expected key strengthens the + forwarding check and matches the requested UI behavior. + +Please approve these adaptations before their results count. The builder +continues static packet preparation only until Sage's test CLEAR. diff --git a/agents/rocko/work/queue-219/base.txt b/agents/rocko/work/queue-219/base.txt new file mode 100644 index 000000000..593df1d00 --- /dev/null +++ b/agents/rocko/work/queue-219/base.txt @@ -0,0 +1 @@ +062646487451993302c768c6812b827c834e952b diff --git a/agents/rocko/work/queue-219/build.patch b/agents/rocko/work/queue-219/build.patch new file mode 100644 index 000000000..9794fce33 --- /dev/null +++ b/agents/rocko/work/queue-219/build.patch @@ -0,0 +1,777 @@ +diff --git a/packages/bus/src/broker.mjs b/packages/bus/src/broker.mjs +index c50f8ac23..5ee242f1d 100644 +--- a/packages/bus/src/broker.mjs ++++ b/packages/bus/src/broker.mjs +@@ -4,7 +4,7 @@ import { initializeControl, processingStatus } from '../../processing-control/sr + import { NAME as HARNESS_NAME } from '../../plugins/src/kinds.mjs'; + import { loadRequirements } from '../../tasks/src/requirements.mjs'; + import { MERGE, LandingRefusal, readLanding } from './landing.mjs'; +-import { ID_PATTERN, THREAD_DEFINITION, END_PROTECTED, forkProblem, forkUrl, credentialKind, validArgv, commandAllowed, HOST_ID, CREATIVE_DEFINITIONS, unhireableAction, giteaName, parseGiteaUrl, covers } from '../../business/src/vocabulary.mjs'; ++import { ID_PATTERN, THREAD_DEFINITIONS, END_PROTECTED, forkProblem, forkUrl, credentialKind, validArgv, commandAllowed, HOST_ID, CREATIVE_DEFINITIONS, unhireableAction, giteaName, parseGiteaUrl, covers } from '../../business/src/vocabulary.mjs'; + + // Kept closed while S1 lands; its vocabulary is the integration authority. + export const ACTIONS = Object.freeze([ +@@ -312,7 +312,7 @@ const enrollBinding = (enroll) => ({ + }); + // r6 E3: the installer line's comment, then the tag's 32 hex digits. + const INSTALLER_COMMENT = 'mosaic-installer-'; +-// Row 121 (#1615): an instance of THREAD_DEFINITION (vocabulary.mjs since row 155) talks with the ++// Row 121 (#1615): an instance of THREAD_DEFINITIONS (vocabulary.mjs since row 213) talks with the + // user in threads. Its runs claim a thread, not the role, and every message to or from it names its + // thread (REQ-MSG-1). + // Row 146 (#1640): the sender of the bus's notice to launch.by. A role instance name matches the +@@ -593,7 +593,7 @@ export class Broker { + ); + } + #threaded(business, role) { +- return this.#business(business).roles[role]?.definition === THREAD_DEFINITION; ++ return THREAD_DEFINITIONS.includes(this.#business(business).roles[role]?.definition); + } + // Row 264a (#1769): a claim holds one session of a thread, and each session has its own holder. + #sessionHolder(session) { +@@ -5790,24 +5790,27 @@ export class Broker { + .all('SELECT id,at,name,repo FROM projects WHERE business=? ORDER BY seq', s.business) + .filter((p) => !s.user || this.#granted(s, 'project.list', p.id, null)); + } +- // A thread belongs to the user who made it and to the business's one thread instance. ++ // A thread belongs to the user who made it and to one declared thread instance. + // Row 177 (#1675): a tester's thread is owned by `tester`, so neither reaches the other's. + case 'thread.create': { + if (!s.tester && !s.user) this.#human(s); +- keys(a, ['project', 'name'], ['project', 'name']); ++ keys(a, ['project', 'name', 'instance'], ['project', 'name']); + if (!name(a.name)) fail('invalid-request'); ++ const instances = Object.keys(b.roles).filter((r) => this.#threaded(s.business, r)); ++ const instance = a.instance === undefined ++ ? (instances.length === 1 ? instances[0] : null) ++ : identifier(a.instance); + // FED R3 (design 3.5): a user's grant must cover the project and the thread instance, checked + // before the project is looked up, so an ungranted id reads the same whether it exists or not. + if (s.user) { +- const threaded = Object.keys(b.roles).filter((r) => this.#threaded(s.business, r)); +- if (threaded.length !== 1 || !this.#granted(s, 'thread.create', identifier(a.project), threaded[0])) fail('access-refused'); ++ if (!instance || !instances.includes(instance) || !this.#granted(s, 'thread.create', identifier(a.project), instance)) ++ fail('access-refused'); + } + if (!this.#store.get('SELECT 1 FROM projects WHERE id=? AND business=?', identifier(a.project), s.business)) + fail('project-not-found'); +- // The bus never picks between instances; the business file declares exactly one. +- const instances = Object.keys(b.roles).filter((r) => this.#threaded(s.business, r)); +- if (!instances.length) fail('thread-instance-missing'); +- if (instances.length > 1) fail('thread-instance-ambiguous'); ++ // The human selects an instance; omission is compatible only with one candidate. ++ if (!instances.length || (a.instance !== undefined && !instances.includes(instance))) fail('thread-instance-missing'); ++ if (a.instance === undefined && instances.length > 1) fail('thread-instance-ambiguous'); + if (this.#store.get('SELECT 1 FROM threads WHERE project=? AND name=?', a.project, a.name)) fail('thread-exists'); + const tid = randomUUID(); + this.#store.run( +@@ -5816,13 +5819,13 @@ export class Broker { + this.#now(), + s.business, + a.project, +- instances[0], ++ instance, + this.#owner(s), + a.name, + ); + // Row 264a (#1769, D3): every thread has a default session from the start. +- this.#newSession(s, { id: tid, business: s.business, instance: instances[0], owner: this.#owner(s) }, 'default', 'default', true); +- return { id: tid, instance: instances[0] }; ++ this.#newSession(s, { id: tid, business: s.business, instance, owner: this.#owner(s) }, 'default', 'default', true); ++ return { id: tid, instance }; + } + // Row 177 (#1675): each lists its own threads. The human's list leaves out the tester's, + // and the tester's lists only its own (row 175 listed it the business human's). +diff --git a/packages/bus/tests/creative.test.mjs b/packages/bus/tests/creative.test.mjs +index 6cfbf1cd6..6d421ba19 100644 +--- a/packages/bus/tests/creative.test.mjs ++++ b/packages/bus/tests/creative.test.mjs +@@ -44,15 +44,18 @@ function setup(t) { + const caps = {}; + caps.human = b.bindHuman({ business: 'demo', human: 'jason', via: 'cli', outsideAgent: true }); + const project = b.request(caps.human, { verb: 'project.create', args: { name: 'demo' } }).id; +- const thread = b.request(caps.human, { verb: 'thread.create', args: { project, name: 'main' } }).id; ++ const thread = b.request(caps.human, { verb: 'thread.create', args: { project, name: 'main', instance: 'assistant' } }).id; ++ const threads = { assistant: thread }; + for (const role of Object.keys(roles)) { +- caps[role] = b.bindLaunch({ business: 'demo', role, run: `${role}-run`, harness: 'pi', ...(role === 'assistant' ? { thread } : {}) }); ++ if (['bean', 'quill'].includes(role)) ++ threads[role] = b.request(caps.human, { verb: 'thread.create', args: { project, name: role, instance: role } }).id; ++ caps[role] = b.bindLaunch({ business: 'demo', role, run: `${role}-run`, harness: 'pi', ...(threads[role] ? { thread: threads[role] } : {}) }); + b.request(caps[role], { verb: 'role.claim' }); + } + caps.tester = b.bindTester({ business: 'demo' }); + const call = (who, verb, args = {}) => b.request(caps[who], { verb, args }); + const count = (sql) => store.get(sql).n; +- return { call, count, store }; ++ return { call, count, store, threads }; + } + + const raise = { action: 'policy.change', question: 'Change?', options: [{ key: 'yes', text: 'Yes' }, { key: 'no', text: 'No' }], recommendation: 'yes', blocking: false }; +@@ -72,7 +75,7 @@ test('M6: a creative raise refuses decision-not-allowed, cited or not, and no de + }); + + test('a creative instance sends only to the user and hears only from the user', (t) => { +- const { call, count } = setup(t); ++ const { call, count, threads } = setup(t); + for (const who of ['bean', 'quill']) { + call(who, 'message.send', { to: 'human', body: 'hello' }); + for (const to of ['chief', 'assistant', 'pm', 'bean', 'quill'].filter((n) => n !== who)) { +@@ -81,7 +84,7 @@ test('a creative instance sends only to the user and hears only from the user', + for (const from of ['chief', 'pm']) { + assert.throws(() => call(from, 'message.send', { to: who, body: 'hello' }), { code: 'message-sender-refused' }, `${from} to ${who}`); + } +- call('human', 'message.send', { to: who, body: 'hello' }); ++ call('human', 'thread.send', { thread: threads[who], body: 'hello' }); + } + assert.equal(count("SELECT count(*) n FROM messages WHERE from_role IN ('bean','quill')"), 2); + assert.equal(count("SELECT count(*) n FROM messages WHERE to_role IN ('bean','quill')"), 2); +diff --git a/packages/bus/tests/threads.test.mjs b/packages/bus/tests/threads.test.mjs +index 8b31da9f9..1410ec3fa 100644 +--- a/packages/bus/tests/threads.test.mjs ++++ b/packages/bus/tests/threads.test.mjs +@@ -601,3 +601,42 @@ test('project.create keeps unrelated names with and without a repository', t => + { id: linked.id, name: 'work', repo: 'stack' }, + ]); + }); ++ ++test('C2-09a: explicit creative instances bind independent threads, default sessions and holders', t => { ++ const creative = ['companion', 'ideas', 'research', 'proofreader', 'fact-checker', 'continuity']; ++ const extra = Object.fromEntries(creative.map((definition, i) => [`writer${i}`, { definition, authority: { withinRole: ['message.send'], crossRole: [] } }])); ++ const ctx = setup(t, demo({ roles: { ...roles, ...extra }, definitions: { ++ ...definitions, ...Object.fromEntries(creative.map(definition => [definition, { messaging: shipped(definition).messaging }])), ++ } })); ++ const project = ctx.call('human', 'project.create', { name: 'Writing' }).id; ++ assert.throws(() => ctx.call('human', 'thread.create', { project, name: 'ambiguous' }), { code: 'thread-instance-ambiguous' }); ++ for (const [instance, role] of Object.entries(extra)) { ++ const receipt = ctx.call('human', 'thread.create', { project, name: instance, instance }); ++ assert.equal(receipt.instance, instance); ++ assert.equal(ctx.store.get('SELECT instance FROM threads WHERE id=?', receipt.id).instance, instance); ++ assert.equal(ctx.store.get('SELECT instance FROM sessions WHERE thread=?', receipt.id).instance, instance); ++ ctx.launch(instance, instance, { thread: receipt.id }); ++ assert.equal(ctx.claim(instance).thread, receipt.id); ++ const sent = ctx.call('human', 'thread.send', { thread: receipt.id, body: `For ${instance}` }); ++ const read = ctx.call(instance, 'message.receive'); ++ assert.deepEqual(read.map(m => m.id), [sent.id], role.definition); ++ assert.equal(ctx.threadOf(sent.id), receipt.id); ++ assert.equal(ctx.store.get('SELECT count(*) n FROM role_claims WHERE role=?', instance).n, 0); ++ assert.throws(() => ctx.call('human', 'message.send', { to: instance, body: 'No implicit thread' }), ++ { code: 'thread-required-use-thread-send' }); ++ } ++ assert.equal(ctx.count('SELECT count(*) n FROM threads'), creative.length); ++}); ++ ++test('C2-09a: explicit missing, non-thread and malformed selectors create no thread or session', t => { ++ const ctx = setup(t); ++ const project = ctx.call('human', 'project.create', { name: 'Writing' }).id; ++ for (const instance of ['not-declared', 'pm', 'coder']) ++ assert.throws(() => ctx.call('human', 'thread.create', { project, name: instance, instance }), { code: 'thread-instance-missing' }); ++ for (const instance of [null, 7, {}, [], '', 'bad name']) ++ assert.throws(() => ctx.call('human', 'thread.create', { project, name: 'bad', instance }), { code: 'invalid-request' }); ++ assert.equal(ctx.count('SELECT count(*) n FROM threads'), 0); ++ assert.equal(ctx.count('SELECT count(*) n FROM sessions'), 0); ++ const legacy = ctx.call('human', 'thread.create', { project, name: 'Default' }); ++ assert.equal(legacy.instance, 'assistant'); ++}); +diff --git a/packages/bus/tests/user-principal.test.mjs b/packages/bus/tests/user-principal.test.mjs +index 4dcd1ebeb..8471ddf85 100644 +--- a/packages/bus/tests/user-principal.test.mjs ++++ b/packages/bus/tests/user-principal.test.mjs +@@ -23,7 +23,7 @@ const grant = (over = {}) => ({ + }); + + // `processing` stands in for the halt switch, as lifecycle-verbs.test.mjs's T15 does. +-function setup(t, { grants = [grant()], processing = null } = {}) { ++function setup(t, { grants = [grant()], processing = null, threadRoles = {} } = {}) { + const root = mkdtempSync(join(tmpdir(), 'bus-user-principal-')); + const store = new Store(root); + t.after(() => { +@@ -40,6 +40,7 @@ function setup(t, { grants = [grant()], processing = null } = {}) { + roles: { + chief: { definition: 'chief-of-staff', authority: { withinRole: ['message.send'], crossRole: [] } }, + assistant: { definition: 'assistant', authority: { withinRole: ['message.send'], crossRole: [] } }, ++ ...threadRoles, + }, + definitions: { 'chief-of-staff': {}, assistant: { messaging: shipped('assistant').messaging } }, + }, +@@ -394,3 +395,21 @@ test('R3 r4: the chief\'s hire refuses an instance named user with the code it g + test('R3 r4: a non-function access throws before the store is read', () => { + assert.throws(() => new Broker({ store: null, businesses: { demo: {} }, access: 'access.json' }), /access must be a function/); + }); ++ ++test('C2-09a: naming an instance cannot bypass its collaborator grant or disclose an ungranted project', t => { ++ const ctx = setup(t, { threadRoles: { bean: { definition: 'companion', authority: { withinRole: ['message.send'], crossRole: [] } } } }); ++ const project = ctx.call('human', 'project.create', { name: 'shared' }).id; ++ const privateProject = ctx.call('human', 'project.create', { name: 'private' }).id; ++ ctx.access.doc.grants = [grant({ projects: [project] })]; ++ ctx.user('alice'); ++ assert.equal(ctx.code('alice', 'thread.create', { project, name: 'blocked bean', instance: 'bean' }), 'access-refused'); ++ for (const id of [privateProject, '00000000-0000-4000-8000-000000000000']) ++ assert.equal(ctx.code('alice', 'thread.create', { project: id, name: 'blocked project', instance: 'assistant' }), 'access-refused'); ++ assert.equal(ctx.store.get('SELECT count(*) n FROM threads').n, 0); ++ const allowed = ctx.call('alice', 'thread.create', { project, name: 'assistant work', instance: 'assistant' }); ++ assert.equal(allowed.instance, 'assistant'); ++ ctx.access.doc.grants = [grant({ projects: [project], instances: ['bean'] })]; ++ const bean = ctx.call('alice', 'thread.create', { project, name: 'bean work', instance: 'bean' }); ++ assert.equal(bean.instance, 'bean'); ++ assert.equal(ctx.code('alice', 'thread.create', { project, name: 'old instance', instance: 'assistant' }), 'access-refused'); ++}); +diff --git a/packages/cli/src/launcher.mjs b/packages/cli/src/launcher.mjs +index e044b7627..f019df83f 100644 +--- a/packages/cli/src/launcher.mjs ++++ b/packages/cli/src/launcher.mjs +@@ -53,7 +53,7 @@ + // session.json names it as gitSocket. + // + // Thread sessions (row 121, #1615). A thread instance (an instance whose +-// role definition is the assistant) runs once per thread, never on its own: ++// role definition is in THREAD_DEFINITIONS) runs once per thread, never on its own: + // a launch from a session or --pm refuses thread-required. Every tick + // (threadTickMs) the host asks the broker for the business's threads + // (launch op threadsState) and: +@@ -139,6 +139,7 @@ import { createServer } from "node:net"; + import { chmodSync, existsSync, lstatSync, mkdirSync, readFileSync, renameSync, rmSync } from "node:fs"; + import { dirname, join, relative } from "node:path"; + import { configDir, loadBusiness, resolveInstance, systemVars } from "../../business/src/index.mjs"; ++import { THREAD_DEFINITIONS } from "../../business/src/vocabulary.mjs"; + import { limitHit } from "../../business/src/launch.mjs"; + import { BundleError, buildBundle, hasFallback, readSkills, sessionModel } from "../../harness/src/bundle.mjs"; + import { createLoader } from "../../plugins/src/loader.mjs"; +@@ -175,8 +176,6 @@ const STOP_TIMEOUT_MS = 30000; + const LINE_MAX = 4096; + // How often the host asks for the threads' state; a poll period, not a limit. + const THREAD_TICK_MS = 2000; +-// The broker's thread instance definition (packages/bus/src/broker.mjs). +-const THREAD_DEFINITION = "assistant"; + // Row 194: the broker's LIFECYCLE_REQUEST_MS. A regenerate's old run must be gone by then. + const LIFECYCLE_REQUEST_MS = 60000; + +@@ -299,7 +298,7 @@ export function createLauncher({ + try { b=loadBusiness(host.business,{dir:configDir(env),rolesDir:rolesDir(env)}); } catch(e){throw new Refusal("organization-refused",e.message);} + if (!b.launch) throw new Refusal("launch-not-configured", `business ${b.id} has no launch block`); + if (!(pm ? instance === b.launch.by : b.launch.instances.includes(instance) || (human && Object.hasOwn(b.roles, instance)))) throw new Refusal("instance-not-listed"); +- const threaded = b.roles[instance]?.definition === THREAD_DEFINITION; ++ const threaded = THREAD_DEFINITIONS.includes(b.roles[instance]?.definition); + if (threaded && !thread) throw new Refusal("thread-required", `${instance} runs only for a thread`); + if (!threaded && thread) throw new Refusal("thread-not-allowed", `${instance} isn't a thread instance`); + if (threaded && !session) throw new Refusal("session-required", `${instance} runs for one session of a thread`); +@@ -854,7 +853,7 @@ export function createLauncher({ + function ending(r) { + const runs = r.runs ?? []; + const at = Date.parse(r.at); +- const threaded = b.roles[r.instance]?.definition === THREAD_DEFINITION; ++ const threaded = THREAD_DEFINITIONS.includes(b.roles[r.instance]?.definition); + return registry.sessions.filter((s) => s.business === b.id && (runs.includes(s.run) || + (!threaded && s.instance === r.instance && (spawnedAt.get(s.run) ?? Date.parse(s.startedAt)) <= at))); + } +diff --git a/packages/cli/tests/launcher.test.mjs b/packages/cli/tests/launcher.test.mjs +index 2202c7753..0a34fe27a 100644 +--- a/packages/cli/tests/launcher.test.mjs ++++ b/packages/cli/tests/launcher.test.mjs +@@ -773,21 +773,28 @@ test("row 213 a creative instance's persona is read from /personas: o + const dir = join(given.root, "config", "personas", "acme"); + mkdirSync(dir, { recursive: true, mode: 0o700 }); + writeFileSync(join(dir, "bean.md"), "Not the pinned text.\n", { mode: 0o600 }); +- const { f, host, launcher } = await setup(t, undefined, given); +- // The chief of staff holds role.launch; a test chief session bound to this process asks. +- const chief = await host.bindLaunch({ business: "acme", role: "chief", run: "chief-test", harness: "pi", pid: process.pid, startTime: startTimeOf(process.pid) }); +- await new Client({ path: host.path, cap: chief.cap }).call("role.claim"); +- // The bundle's refusal comes back as launch-failed, as webPlan's does, with its reason in the log. +- assert.deepEqual(await launcher.fromSession(chief.cap, "bean"), { ok: false, error: "launch-failed" }); ++ // Row 219: creative instances launch from their thread, not from a chief's standalone request. ++ // The mismatch occurs before bindLaunch, so only the host's thread-state snapshot is a fixture. ++ const thread = 'persona-thread'; ++ const { f, launcher } = await setup(t, undefined, given, { wrap: host => ({ ++ ...host, ++ op: message => message.op === 'threadsState' ++ ? Promise.resolve([{ thread, instance: 'bean', holder: null, pending: 1 }]) ++ : host.op(message), ++ }) }); ++ // The bundle's refusal remains launch-failed with the persona mismatch as its reason. ++ await launcher.tick(); + const [e] = log(f); + assert.deepEqual([e.event, e.code, log(f).length], ["refused", "launch-failed", 1]); ++ assert.equal(e.thread, thread); + assert.equal(e.why, `instance bean: persona file ${join(dir, "bean.md")} isn't the file persona.sha256 names`); + assert.ok(!JSON.stringify(log(f)).includes("PERSONA-MARKER-d2")); + assert.deepEqual(readSessions(f.dataRoot), []); + // The run directory exists by then (BUILD.md residual); the persona is read before any bundle file. + const runs = readdirSync(join(f.dataRoot, "launches", "acme")); + assert.equal(runs.length, 1); +- assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), ["session"]); ++ assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), []); ++ assert.equal(existsSync(join(f.dataRoot, "threads", "acme", thread, "session")), true); + }); + + test("row 64 on Claude Code a role skill that isn't enabled refuses the same way, before anything starts", { timeout: 30000 }, async (t) => { +diff --git a/packages/cli/tests/thread-launch.test.mjs b/packages/cli/tests/thread-launch.test.mjs +index 0d207a0a0..84f8ed088 100644 +--- a/packages/cli/tests/thread-launch.test.mjs ++++ b/packages/cli/tests/thread-launch.test.mjs +@@ -396,6 +396,22 @@ test("REQ-THREAD-1 REQ-LAUNCH-1 the chief can't launch the assistant outside a t + assert.deepEqual(launchLog(f).map((e) => [e.event, e.instance, e.code]), [["refused", "assistant", "thread-required"]]); + }); + ++// Admission refuses before bundle construction or a sandbox is needed. ++test('C2-09a: fromSession refuses a creative instance outside a thread without a sandbox', { timeout: 30000 }, async t => { ++ const prepared = prepare(t, { edit: doc => { ++ doc.roles.bean = { ...doc.roles.assistant, definition: 'companion' }; ++ doc.launch.instances.push('bean'); ++ return doc; ++ } }); ++ const { f } = prepared; ++ const host = await start(t, prepared); ++ const chief = await host.host.bindLaunch({ business: 'acme', role: 'chief', run: 'chief-bean-refusal', harness: 'pi', pid: process.pid, startTime: startTimeOf(process.pid) }); ++ await new Client({ path: host.host.path, cap: chief.cap }).call('role.claim'); ++ assert.deepEqual(await host.launcher.fromSession(chief.cap, 'bean'), { ok: false, error: 'thread-required' }); ++ assert.deepEqual(readSessions(f.dataRoot), []); ++ assert.deepEqual(launchLog(f).map(e => [e.event, e.instance, e.code]), [['refused', 'bean', 'thread-required']]); ++}); ++ + // Row 125 (#1619, REQ-LIMIT-3, REQ-LIMIT-4): the token-limit check switched on, with the + // assistant's limit at 0, so it is at its limit before any turn. The real check runs through the + // bus's plug-in loader. +@@ -443,3 +459,71 @@ test("REQ-LIMIT-4 a thread run at its token limit stops at the end of its turn: + assert.deepEqual(threadRuns(f, A), []); + assert.equal(host.runtime.threadsState("acme").find((x) => x.thread === A).holder, null); + }); ++ ++test('C2-09a: creative thread messages start distinct runs and keep each instance history separate', { skip, timeout: 90000 }, async t => { ++ const prepared = prepare(t, { sonnet: 2, idle: 60, edit: doc => { ++ const vars = { ...doc.roles.assistant.vars }; ++ delete doc.roles.assistant; ++ doc.roles.bean = { definition: 'companion', vars: { ...vars } }; ++ doc.roles.chet = { definition: 'continuity', vars: { ...vars } }; ++ doc.launch.instances = ['coder', 'bean', 'chet']; ++ return doc; ++ } }); ++ const host = await start(t, prepared), { f } = prepared; ++ const project = host.call('project.create', { name: 'Writing' }).id; ++ const bean = host.call('thread.create', { project, name: 'Bean', instance: 'bean' }).id; ++ const chet = host.call('thread.create', { project, name: 'Chet', instance: 'chet' }).id; ++ const send = async (thread, instance, body) => { ++ const message = host.call('thread.send', { thread, body }); ++ const outcome = await until(() => host.call('thread.messages', { thread }).find(m => m.in_reply_to === message.id) || ++ launchLog(f).find(row => row.thread === thread && row.event === 'refused')); ++ assert.notEqual(outcome.event, 'refused', `creative thread launch refused: ${outcome.code}`); ++ const reply = outcome; ++ assert.equal(reply.from_role, instance); ++ assert.equal(reply.to_role, 'human'); ++ return JSON.parse(reply.body.match(/THREAD (\{.*\})/)[1]); ++ }; ++ assert.deepEqual((await send(bean, 'bean', 'Bean input')).before, []); ++ assert.deepEqual((await send(chet, 'chet', 'Chet input')).before, []); ++ assert.deepEqual((await send(bean, 'bean', 'Bean again')).before, ['Bean input']); ++ assert.equal(readFileSync(join(f.dataRoot, 'threads', 'acme', bean, 'session', 'history'), 'utf8'), 'Bean input\nBean again\n'); ++ assert.equal(readFileSync(join(f.dataRoot, 'threads', 'acme', chet, 'session', 'history'), 'utf8'), 'Chet input\n'); ++ const launches = launchLog(f).filter(row => row.event === 'launch'); ++ assert.ok(launches.some(row => row.instance === 'bean' && row.thread === bean)); ++ assert.ok(launches.some(row => row.instance === 'chet' && row.thread === chet)); ++ const chief = await host.host.bindLaunch({ business: 'acme', role: 'chief', run: 'chief-creative-test', harness: 'pi', pid: process.pid, startTime: startTimeOf(process.pid) }); ++ await new Client({ path: host.host.path, cap: chief.cap }).call('role.claim'); ++ assert.deepEqual(await host.launcher.fromSession(chief.cap, 'bean'), { ok: false, error: 'thread-required' }); ++}); ++ ++test('C2-09a: terminating one creative thread does not signal an unselected peer of the same instance', { skip, timeout: 90000 }, async t => { ++ const prepared = prepare(t, { sonnet: 2, idle: 60, edit: doc => { ++ doc.roles.bean = { ...doc.roles.assistant, definition: 'companion' }; ++ delete doc.roles.assistant; ++ doc.launch.instances = ['coder', 'bean']; ++ return doc; ++ } }); ++ let requests = []; ++ const host = await start(t, prepared, { ops: () => ({ ++ lifecycleRequests: () => requests, ++ lifecycleTake: () => ({ ...requests[0], go: true }), ++ lifecycleDone: () => { requests = []; return {}; }, ++ }) }); ++ const { f } = prepared; ++ const project = host.call('project.create', { name: 'Writing' }).id; ++ const threads = ['one', 'two'].map(name => host.call('thread.create', { project, name, instance: 'bean' }).id); ++ for (const thread of threads) { ++ const message = host.call('thread.send', { thread, body: 'Stay in this thread' }); ++ const outcome = await until(() => host.call('thread.messages', { thread }).find(row => row.in_reply_to === message.id) || ++ launchLog(f).find(row => row.thread === thread && row.event === 'refused')); ++ assert.notEqual(outcome.event, 'refused', `creative thread launch refused: ${outcome.code}`); ++ } ++ const runs = threads.map(thread => launchLog(f).find(row => row.thread === thread && row.event === 'launch').run); ++ requests = [{ request: 'one-thread-end', state: 'open', business: 'acme', action: 'lifecycle.end', ++ instance: 'bean', runs: [runs[0]], by: f.doc.human, at: new Date(Date.now() + 1000).toISOString() }]; ++ await host.launcher.tick(); ++ await until(() => launchLog(f).some(row => row.event === 'lifecycle-end' && row.run === runs[0])); ++ assert.deepEqual(launchLog(f).filter(row => row.event === 'lifecycle-end').map(row => row.run), [runs[0]]); ++ await until(() => !readSessions(f.dataRoot).some(row => row.run === runs[0])); ++ assert.ok(readSessions(f.dataRoot).some(row => row.run === runs[1])); ++}); +diff --git a/packages/webui/src/chat-http.mjs b/packages/webui/src/chat-http.mjs +index c7770726d..396b42f3c 100644 +--- a/packages/webui/src/chat-http.mjs ++++ b/packages/webui/src/chat-http.mjs +@@ -78,7 +78,8 @@ export async function chatHTTP(req, res, { path, search, chat, named, send, prev + (Object.hasOwn(v, 'repo') && (typeof v.repo !== 'string' || !KEY.test(v.repo)))) throw Error(); + action = 'createProject'; args = v; + } else if (path === '/api/chat/threads') { +- if (!exact(v, ['project', 'name']) || !id(v.project) || !name(v.name)) throw Error(); ++ if (!exact(v, ['project', 'name', 'instance'], ['project', 'name']) || !id(v.project) || !name(v.name) || ++ (Object.hasOwn(v, 'instance') && (typeof v.instance !== 'string' || !KEY.test(v.instance)))) throw Error(); + action = 'createThread'; args = v; + } else if (match?.[2] === 'messages') { + if (!exact(v, ['body']) || typeof v.body !== 'string' || !v.body.trim() || v.body.length > 32768) throw Error(); +diff --git a/packages/webui/src/chat.mjs b/packages/webui/src/chat.mjs +index 018574ad4..65b89dab3 100644 +--- a/packages/webui/src/chat.mjs ++++ b/packages/webui/src/chat.mjs +@@ -20,7 +20,7 @@ export function createChat({ call }) { + projects: () => channel('project.list', {}), + createProject: ({ name, repo }) => channel('project.create', { name, ...(repo === undefined ? {} : { repo }) }), + threads: ({ project } = {}) => channel('thread.list', project === undefined ? {} : { project }), +- createThread: ({ project, name }) => channel('thread.create', { project, name }), ++ createThread: ({ project, name, instance }) => channel('thread.create', { project, name, ...(instance === undefined ? {} : { instance }) }), + messages: ({ thread, after, limit }) => channel('thread.messages', { thread, ...(after ? { after } : {}), limit }), + send: ({ thread, body }) => channel('thread.send', { thread, body, class: 'REQUEST' }), + open: ({ thread }) => channel('thread.open', { thread }), +diff --git a/packages/webui/src/public/chat.js b/packages/webui/src/public/chat.js +index a6b805c1b..15066bdc6 100644 +--- a/packages/webui/src/public/chat.js ++++ b/packages/webui/src/public/chat.js +@@ -14,7 +14,7 @@ + return states.get(id); + }; + const form = key => { +- if (!forms.has(key)) forms.set(key, { name: '', repo: '', receipt: '', pending: false }); ++ if (!forms.has(key)) forms.set(key, { name: '', repo: '', instance: '', instanceChosen: false, receipt: '', pending: false }); + return forms.get(key); + }; + class ChatError extends Error { +@@ -58,7 +58,11 @@ + function creation(kind, key) { + const f = form(key), project = kind === 'project'; + const repo = project && current.repos.length ? `` : ''; +- return `

New ${kind}

${repo}

${txt(f.receipt)}

`; ++ const choices = current.instances ?? []; ++ const picker = project ? '' : ``; ++ const unavailable = !project && (!!current.instanceError || !choices.some(row => row.instance === f.instance)); ++ const status = !project && current.instanceError ? current.instanceError : f.receipt || (!project && !choices.length ? 'No thread instances are configured.' : ''); ++ return `

New ${kind}

${repo}${picker}

${txt(status)}

`; + } + function threads(rows) { + return rows.length ? `` : '

No threads yet.

'; +@@ -106,7 +110,7 @@ + const f = document.getElementById(keep.field.id); + if (f && !f.disabled) { + f.focus({ preventScroll: true }); +- if (keep.field.start !== null) f.setSelectionRange(keep.field.start, keep.field.end); ++ if (keep.field.start != null && typeof f.setSelectionRange === 'function') f.setSelectionRange(keep.field.start, keep.field.end); + } + } else if (navigated) c.container.querySelector('h1')?.focus({ preventScroll: true }); + } +@@ -126,7 +130,7 @@ + paint(); return; + } + const old = current; +- current = { container, kind, id, canSend, projects: old?.projects ?? [], threads: old?.threads ?? [], repos: old?.repos ?? [], error: '', refused: false }; ++ current = { container, kind, id, canSend, projects: old?.projects ?? [], threads: old?.threads ?? [], repos: old?.repos ?? [], instances: old?.instances ?? [], instanceError: kind === 'project' ? 'Reading thread instances…' : '', error: '', refused: false }; + if (navigated || !old) container.innerHTML = '

Reading chat…

'; + try { + const s = kind === 'thread' ? state(id) : null; +@@ -147,6 +151,25 @@ + s.receipt = 'Message shown in this conversation.'; s.receiptId = null; + } + } ++ if (kind === 'project') { ++ try { ++ const listed = await request('instances', undefined, signal); ++ if (g !== generation || out()) return; ++ if (!listed.rows.every(row => typeof row?.instance === 'string' && /^[a-zA-Z0-9][a-zA-Z0-9_.-]{0,159}$/.test(row.instance) && typeof row.name === 'string')) ++ throw new ChatError('invalid-response'); ++ current.instances = listed.rows; current.instanceError = ''; ++ const f = form(id); ++ if (f.instance && !listed.rows.some(row => row.instance === f.instance)) { ++ f.instance = ''; f.instanceChosen = true; ++ f.receipt = 'The selected instance is no longer available. Choose who this thread is with.'; ++ } ++ if (!f.instanceChosen && !f.instance && listed.rows.length === 1) f.instance = listed.rows[0].instance; ++ } catch (err) { ++ if (err.name === 'AbortError' || out() || g !== generation) return; ++ current.instances = []; ++ current.instanceError = "Couldn't read the thread instances; reload to try again"; ++ } ++ } + // Repository choices are optional metadata; an unavailable read must not block unlinked chat. + if (kind === 'list') { + try { +@@ -179,7 +202,11 @@ + if (!current || out()) return; + if (event.target.id === 'chat-body') state(current.id).draft = event.target.value; + const f = event.target.closest('.chat-create'); +- if (f && event.target.dataset.chatField) form(f.dataset.key)[event.target.dataset.chatField] = event.target.value; ++ if (f && event.target.dataset.chatField) { ++ const draft = form(f.dataset.key), field = event.target.dataset.chatField; ++ draft[field] = event.target.value; ++ if (field === 'instance') { draft.instanceChosen = true; draft.receipt = ''; paint(); } ++ } + }); + document.addEventListener('scroll', event => { + if (event.target.matches?.('.chat-log') && current?.kind === 'thread') { +@@ -232,7 +259,8 @@ + void write('send', id, { body: state(id).draft }, `threads/${id}/messages`); + } else if (event.target.classList.contains('chat-create')) { + const key = event.target.dataset.key, f = form(key), project = key === 'projects'; +- const value = project ? { name: f.name, ...(f.repo ? { repo: f.repo } : {}) } : { project: key, name: f.name }; ++ if (!project && (current.instanceError || !(current.instances ?? []).some(row => row.instance === f.instance))) return; ++ const value = project ? { name: f.name, ...(f.repo ? { repo: f.repo } : {}) } : { project: key, name: f.name, instance: f.instance }; + void write(project ? 'createProject' : 'createThread', key, value, project ? 'projects' : 'threads'); + } + }); +diff --git a/packages/webui/src/serve.mjs b/packages/webui/src/serve.mjs +index 288871d0c..d9d27a062 100644 +--- a/packages/webui/src/serve.mjs ++++ b/packages/webui/src/serve.mjs +@@ -8,6 +8,7 @@ import { BusReadError, busStatus as coreBusStatus, subjectOk, sendOutcome } from + // Row 176, sage-1111: HTTP-only tester refusals; the row-164 transport pin stays untouched. + const busStatus = code => ['tester-off', 'tester-required'].includes(code) ? 403 : coreBusStatus(code); + import { ReadError, readStatus, fieldTable } from './reads.mjs'; ++import { THREAD_DEFINITIONS } from '../../business/src/vocabulary.mjs'; + import { chatHTTP } from './chat-http.mjs'; + import { processingHTTP } from './processing-http.mjs'; + +@@ -705,6 +706,25 @@ export async function startServer({ host = '127.0.0.1', port = 7330, board = DEF + if (req.method !== 'POST') return json(res, 405, { error: 'method not allowed' }); + return busSend(req, res, { send, preview: signer, log }); + } ++ // Row 219: labels only, from this listener's own human or tester bus, after auth above. ++ if (path === '/api/chat/instances') { ++ if (req.method !== 'GET') return json(res, 405, { error: 'method-not-allowed' }); ++ if (search) return json(res, 400, { error: 'invalid-request' }); ++ if (!bus?.instances) return json(res, 503, { error: 'not-configured' }); ++ try { ++ const value = await bus.instances(); ++ if (!value || !value.instances || typeof value.instances !== 'object' || Array.isArray(value.instances)) ++ throw new BusReadError('invalid-response'); ++ const rows = Object.entries(value.instances) ++ .filter(([, entry]) => THREAD_DEFINITIONS.includes(entry?.definition)) ++ .map(([instance, entry]) => ({ instance, name: entry.vars?.['persona.name'] || instance })); ++ if (rows.some(row => typeof row.name !== 'string')) throw new BusReadError('invalid-response'); ++ return json(res, 200, { rows, at: new Date().toISOString() }); ++ } catch (err) { ++ const code = err instanceof BusReadError ? err.code : 'invalid-response'; ++ return json(res, busStatus(code), { error: code }); ++ } ++ } + if (path === '/api/chat' || path.startsWith('/api/chat/')) { + return chatHTTP(req, res, { path, search, chat, named, send, preview: writer, log }); + } +diff --git a/packages/webui/tests/chat-browser.test.mjs b/packages/webui/tests/chat-browser.test.mjs +index f3eeeca96..cc3ec093d 100644 +--- a/packages/webui/tests/chat-browser.test.mjs ++++ b/packages/webui/tests/chat-browser.test.mjs +@@ -12,6 +12,7 @@ const HOSTILE = '\u001b[31m\u202eevil'; + const stub = `(() => { + const real = fetch.bind(window); + window.__writes=[];window.__reads=[];window.__answers=[];window.__pages=[]; ++ window.__instances=[{instance:'assistant',name:'assistant'}]; + window.__projects=[{id:'${P}', name:${JSON.stringify(HOSTILE)}, repo:null},{id:'\">',name:'Invalid record'}]; + window.__threads=[{id:'${A}',project:'${P}',name:'alpha',instance:'assistant'}, {id:'${B}',project:'${P}',name:'beta',instance:'assistant'}]; + window.fetch=async (url,init={})=>{ +@@ -33,6 +34,7 @@ const stub = `(() => { + window.__reads.push(path); + if(window.__holdRead && path.includes(window.__holdRead)){window.__heldSignal=init.signal;await new Promise(r=>window.__releaseRead=r);} + if(window.__readFailure)return new Response(JSON.stringify({error:window.__readFailure}),{status:window.__readFailure==='thread-not-owned'?403:503}); ++ if(path==='/api/chat/instances')return window.__instanceFailure ? new Response(JSON.stringify({error:'not-configured'}),{status:503}) : data({rows:window.__instances}); + if(path==='/api/chat/projects')return data({rows:window.__projects}); + if(path.startsWith('/api/chat/threads')&&!path.includes('/messages'))return data({rows:window.__threads}); + if(window.__pages.length)return data(window.__pages.shift()); +@@ -82,7 +84,7 @@ test('chat browser: list/create named projects and threads, hostile labels inert + await b.evaluate(`window.__answers.push({body:{id:'${A}',instance:'assistant'}}),true`); + await click('#chat-thread-form button[type=submit]'); + await wait('document.querySelector("#chat-send-form")'); +- assert.deepEqual(await b.evaluate('window.__writes[1].body'), { project: P, name: 'named thread' }); ++ assert.deepEqual(await b.evaluate('window.__writes[1].body'), { project: P, name: 'named thread', instance: 'assistant' }); + assert.equal(await b.evaluate('window.__writes[1].path'), '/api/chat/threads'); + assert.equal(await b.evaluate('window.__writes.length'), 2); + }); +@@ -264,3 +266,65 @@ test('chat browser: a late project creation receipt cannot navigate away from an + assert.equal(await b.evaluate('location.hash'), '#/chat/threads/' + A); + assert.equal(await b.evaluate('document.querySelector("#chat-body").value'), 'stay in alpha'); + }); ++ ++test('C2-09a picker: no silent choice, duplicate hostile labels stay inert, selection survives refresh and reaches POST', async t => { ++ const { b, wait, set, click, refresh } = await setup(t); ++ await b.evaluate(`window.__instances=[{instance:'bean',name:${JSON.stringify(HOSTILE)}},{instance:'chet',name:${JSON.stringify(HOSTILE)}}];location.hash='#/chat/projects/${P}';true`); ++ await wait('document.querySelectorAll("#chat-instance option").length===3'); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), ''); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").required'), true); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); ++ const labels = await b.evaluate('[...document.querySelectorAll("#chat-instance option")].map(o=>o.textContent)'); ++ assert.match(labels[1], /\(bean\)/); ++ assert.match(labels[2], /\(chet\)/); ++ assert.doesNotMatch(labels.join(''), /[\u202a-\u202e\u2066-\u2069]/u); ++ assert.equal(await b.evaluate('!!document.querySelector("#chat-thread-form img")'), false); ++ assert.equal(await b.evaluate('window.injected'), undefined); ++ await set('#chat-thread-name', 'Bean discussion'); ++ await click('#chat-thread-form button'); ++ assert.equal(await b.evaluate('window.__writes.length'), 0); ++ await set('#chat-instance', 'bean'); ++ await refresh(); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), 'bean'); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-name").value'), 'Bean discussion'); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), false); ++ await b.evaluate(`window.__answers.push({body:{id:'${A}',instance:'bean'}});true`); ++ await click('#chat-thread-form button'); ++ await wait('window.__writes.length===1'); ++ assert.deepEqual(await b.evaluate('window.__writes[0]'), { path: '/api/chat/threads', body: { project: P, name: 'Bean discussion', instance: 'bean' }, redirect: 'manual' }); ++}); ++ ++test('C2-09a picker: metadata failure, empty set and removed pick explain disabled state and recover without switching', async t => { ++ const { b, wait, set, refresh } = await setup(t); ++ await b.evaluate(`window.__instanceFailure=true;location.hash='#/chat/projects/${P}';true`); ++ await wait('document.querySelector("#chat-thread-form") && !document.querySelector(".chat-loading")'); ++ assert.match(await b.evaluate('document.querySelector("#chat-thread-form [role=status]").textContent'), /read the thread instances/); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); ++ await b.evaluate('window.__instanceFailure=false;window.__instances=[];true'); ++ await refresh(); ++ assert.match(await b.evaluate('document.querySelector("#chat-thread-form [role=status]").textContent'), /No thread instances/); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); ++ await b.evaluate("window.__instances=[{instance:'bean',name:'Bean'},{instance:'chet',name:'Chet'}];true"); ++ await refresh(); ++ await set('#chat-instance', 'bean'); ++ await b.evaluate("window.__instances=[{instance:'chet',name:'Chet'}];true"); ++ await refresh(); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), ''); ++ assert.match(await b.evaluate('document.querySelector("#chat-thread-form [role=status]").textContent'), /no longer available/); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); ++ await refresh(); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), '', 'another refresh must not silently select the one remaining instance'); ++ await set('#chat-instance', 'chet'); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), false); ++}); ++ ++test('C2-09a picker: a single instance is chosen explicitly but read-only mode cannot create a thread', async t => { ++ const { b, wait } = await setup(t, '?readonly'); ++ await b.evaluate(`location.hash='#/chat/projects/${P}';true`); ++ await wait('document.querySelectorAll("#chat-instance option").length===2'); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), 'assistant'); ++ assert.equal(await b.evaluate('document.querySelector("#chat-instance").disabled'), true); ++ assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); ++ assert.match(await b.evaluate('document.querySelector(".chat-area").textContent'), /only reads/); ++ assert.equal(await b.evaluate('window.__writes.length'), 0); ++}); +diff --git a/packages/webui/tests/chat-instances.test.mjs b/packages/webui/tests/chat-instances.test.mjs +new file mode 100644 +index 000000000..893c38f80 +--- /dev/null ++++ b/packages/webui/tests/chat-instances.test.mjs +@@ -0,0 +1,87 @@ ++import test from 'node:test'; ++import assert from 'node:assert/strict'; ++import { request } from 'node:http'; ++import { mkdtempSync, rmSync } from 'node:fs'; ++import { join } from 'node:path'; ++import { tmpdir } from 'node:os'; ++import { startServer } from '../src/serve.mjs'; ++import { startTesterSocket, TESTER_ORIGIN } from '../src/tester-socket.mjs'; ++import { close } from './fixture.mjs'; ++ ++const path = '/api/chat/instances'; ++const preview = { host: 'chat.example', header: 'x-user', user: 'jason' }; ++const snapshot = { ++ revision: 'a'.repeat(64), plugins: { private: 'do-not-project' }, ++ instances: { ++ assistant: { definition: 'assistant', vars: { model: 'private-model' } }, ++ bean: { definition: 'companion', vars: { 'persona.name': 'Bean', 'persona.sha256': 'private-hash' } }, ++ chet: { definition: 'continuity', vars: { 'persona.name': 'Bean', file: '/private/path' } }, ++ pm: { definition: 'pm', vars: { 'persona.name': 'Hidden PM' } }, ++ }, ++}; ++function get(server, { route = path, method = 'GET', headers = {} } = {}) { ++ const address = server.address(), unix = typeof address === 'string'; ++ return new Promise((resolve, reject) => { ++ const req = request({ ...(unix ? { socketPath: address } : { host: '127.0.0.1', port: address.port }), ++ path: route, method, headers: { host: unix ? new URL(TESTER_ORIGIN).host : preview.host, ...headers } }, res => { ++ let text = ''; ++ res.on('data', chunk => text += chunk); ++ res.on('end', () => resolve({ status: res.statusCode, text, value: JSON.parse(text) })); ++ }); ++ req.on('error', reject); req.end(); ++ }); ++} ++test('C2-09a metadata uses the human listener bus, projects labels only and refuses unauthenticated reads', async t => { ++ let calls = 0; ++ const server = await startServer({ port: 0, preview, bus: { instances: async () => { calls++; return snapshot; } } }); ++ t.after(() => close(server)); ++ const refused = await get(server); ++ assert.equal(refused.status, 401); ++ assert.equal(calls, 0); ++ const answer = await get(server, { headers: { 'x-user': 'jason' } }); ++ assert.equal(answer.status, 200); ++ assert.deepEqual(answer.value.rows, [{ instance: 'assistant', name: 'assistant' }, { instance: 'bean', name: 'Bean' }, { instance: 'chet', name: 'Bean' }]); ++ for (const row of answer.value.rows) assert.deepEqual(Object.keys(row).sort(), ['instance', 'name']); ++ assert.doesNotMatch(answer.text, /private|Hidden PM/); ++ assert.equal(calls, 1); ++ for (const options of [{ route: path + '?extra=1' }, { method: 'POST' }]) { ++ const result = await get(server, { ...options, headers: { 'x-user': 'jason' } }); ++ assert.ok([400, 405].includes(result.status)); ++ } ++ assert.equal(calls, 1, 'invalid metadata requests never call a bus'); ++}); ++test('C2-09a metadata uses the private tester listener own bus and no human transport', async t => { ++ const root = mkdtempSync(join(tmpdir(), 'chat219-')); t.after(() => rmSync(root, { recursive: true, force: true })); ++ let reads = 0; ++ const server = await startTesterSocket({ dataRoot: root, board: 'http://127.0.0.1:9', ++ bus: { instances: async () => { reads++; return { instances: { quill: { definition: 'proofreader', vars: { 'persona.name': 'Quill' } } } }; } } }); ++ t.after(() => close(server)); ++ const answer = await get(server); ++ assert.equal(answer.status, 200); ++ assert.deepEqual(answer.value.rows, [{ instance: 'quill', name: 'Quill' }]); ++ assert.equal(reads, 1); ++}); ++test('C2-09a metadata refuses missing or malformed reads and never echoes untrusted fields', async t => { ++ for (const bus of [null, { instances: async () => ({ instances: [] }) }, { instances: async () => { throw new Error('private-path'); } }]) { ++ const server = await startServer({ port: 0, preview, bus }); t.after(() => close(server)); ++ const answer = await get(server, { headers: { 'x-user': 'jason' } }); ++ assert.ok(answer.status >= 400); ++ assert.doesNotMatch(answer.text, /private-path/); ++ } ++}); ++ ++test('C2-09a metadata stays behind the SSO session check and a GET never refreshes session activity', async t => { ++ let session = null, reads = 0; ++ const touches = []; ++ const sso = { publicHost: 'sso-chat.example', publicOrigin: 'https://sso-chat.example', formTargets: () => [], ++ session: (_req, { touch }) => { touches.push(touch); return session; } }; ++ const server = await startServer({ port: 0, sso, bus: { instances: async () => { reads++; return snapshot; } } }); ++ t.after(() => close(server)); ++ const headers = { host: sso.publicHost }; ++ assert.equal((await get(server, { headers })).status, 401); ++ assert.equal(reads, 0); ++ session = { name: 'jason' }; ++ assert.equal((await get(server, { headers })).status, 200); ++ assert.equal(reads, 1); ++ assert.deepEqual(touches, [false, false]); ++}); +diff --git a/packages/webui/tests/chat.test.mjs b/packages/webui/tests/chat.test.mjs +index 8c11611cf..b47ecf1a7 100644 +--- a/packages/webui/tests/chat.test.mjs ++++ b/packages/webui/tests/chat.test.mjs +@@ -335,3 +335,21 @@ test('the chat seam names the transport it is given: the human one and the teste + assert.throws(() => createChat({}), { name: 'TypeError', message: 'chat needs a bus transport function' }); + assert.equal(typeof createChat({ call: async () => ({}) }).projects, 'function'); + }); ++ ++test('C2-09a createThread forwards an explicit instance through the existing channel', async t => { ++ const c = await setup(t); ++ c.state.answer = { id: T, instance: 'bean' }; ++ const answer = await c.get('/api/chat/threads', { method: 'POST', value: { project: P, name: 'Bean work', instance: 'bean' } }); ++ assert.equal(answer.status, 200); ++ assert.deepEqual(c.calls, [['thread.create', { project: P, name: 'Bean work', instance: 'bean' }]]); ++ assert.equal(answer.data.instance, 'bean'); ++}); ++test('C2-09a malformed instance and extra thread fields refuse before the channel write', async t => { ++ const c = await setup(t); ++ for (const value of [null, 7, {}, [], '', 'bad name', 'bean/../../other']) { ++ const answer = await c.get('/api/chat/threads', { method: 'POST', value: { project: P, name: 'work', instance: value } }); ++ assert.equal(answer.status, 400, JSON.stringify(value)); ++ } ++ assert.equal((await c.get('/api/chat/threads', { method: 'POST', value: { project: P, name: 'work', instance: 'bean', extra: true } })).status, 400); ++ assert.deepEqual(c.calls, []); ++}); diff --git a/agents/rocko/work/queue-219/candidate-manifest.sha256 b/agents/rocko/work/queue-219/candidate-manifest.sha256 new file mode 100644 index 000000000..562f298c7 --- /dev/null +++ b/agents/rocko/work/queue-219/candidate-manifest.sha256 @@ -0,0 +1,14 @@ +38c4cb694cb756cfc771395afbeaa0779737bb94bc55bd1ea3a72a00d5448041 packages/bus/src/broker.mjs +b0c66c8999b567896923f6c876911dba936c6eca1ccc7aaeada51369fd613594 packages/bus/tests/creative.test.mjs +de4913f50801a11fd172e3f2c2b3cb41ab04e84d32a6650d1c8662b5e227ea3d packages/bus/tests/threads.test.mjs +8f7d6745fffb2a0898ac80c012084ac48964748da6439a8c56facc6529ffb19b packages/bus/tests/user-principal.test.mjs +c905590032b6f8eb7d607d04ce6a585ed783a0c84033dd8436eab7922f2251e1 packages/cli/src/launcher.mjs +fc301cde49e6db1286e90d6ac3410f5f4d87351bec61f1bae27c235a041848b6 packages/cli/tests/launcher.test.mjs +a219036737419870e0aa9d0da20843759a66be5bafff44d3292593833f8fcb9a packages/cli/tests/thread-launch.test.mjs +cbfd54e6b375561d1e4679ee8d78a3285f970d9a7198cec9d22a06510a36afba packages/webui/src/chat-http.mjs +f49a49c6ec69783190d3f5eaf34316cb4d2c44d10d2ee9336fe1383207afefff packages/webui/src/chat.mjs +8e89df5fe8a9da35de41b5968afc6efcd92503e92ef28c6f557bb5371d10d1a4 packages/webui/src/public/chat.js +a597cc4902a873fe15650dacfce14dbc897db26310a7b6a24c293287add946c8 packages/webui/src/serve.mjs +d918e5b061ced9db28337a9b02c7042f9ddafaf09f32affce9f51190c0cba93f packages/webui/tests/chat-browser.test.mjs +2bad14535372488de326b82b791b79a61e1a4ab839458535132958ef85b746eb packages/webui/tests/chat-instances.test.mjs +7779d33d5b1dd3fe183722d0f91ff1f71e741e113cf2d5ee407f3d26428fc162 packages/webui/tests/chat.test.mjs diff --git a/agents/rocko/work/queue-219/changed-paths.txt b/agents/rocko/work/queue-219/changed-paths.txt new file mode 100644 index 000000000..689c1cf8e --- /dev/null +++ b/agents/rocko/work/queue-219/changed-paths.txt @@ -0,0 +1,14 @@ +packages/bus/src/broker.mjs +packages/bus/tests/creative.test.mjs +packages/bus/tests/threads.test.mjs +packages/bus/tests/user-principal.test.mjs +packages/cli/src/launcher.mjs +packages/cli/tests/launcher.test.mjs +packages/cli/tests/thread-launch.test.mjs +packages/webui/src/chat-http.mjs +packages/webui/src/chat.mjs +packages/webui/src/public/chat.js +packages/webui/src/serve.mjs +packages/webui/tests/chat-browser.test.mjs +packages/webui/tests/chat-instances.test.mjs +packages/webui/tests/chat.test.mjs diff --git a/agents/rocko/work/queue-219/mutants.json b/agents/rocko/work/queue-219/mutants.json new file mode 100644 index 000000000..10d146810 --- /dev/null +++ b/agents/rocko/work/queue-219/mutants.json @@ -0,0 +1,193 @@ +[ + { + "id": "T01", + "file": "packages/bus/src/broker.mjs", + "old": "return THREAD_DEFINITIONS.includes(this.#business(business).roles[role]?.definition);", + "new": "return this.#business(business).roles[role]?.definition === 'assistant';", + "test": "packages/bus/tests/threads.test.mjs", + "pattern": "C2-09a: explicit creative" + }, + { + "id": "T02", + "file": "packages/bus/src/broker.mjs", + "old": ": identifier(a.instance);", + "new": ": instances[0];", + "test": "packages/bus/tests/threads.test.mjs", + "pattern": "C2-09a: explicit creative" + }, + { + "id": "T03", + "file": "packages/bus/src/broker.mjs", + "old": "if (!instances.length || (a.instance !== undefined && !instances.includes(instance))) fail('thread-instance-missing');", + "new": "if (!instances.length) fail('thread-instance-missing');", + "test": "packages/bus/tests/threads.test.mjs", + "pattern": "C2-09a: explicit missing" + }, + { + "id": "T04", + "file": "packages/bus/src/broker.mjs", + "old": "if (a.instance === undefined && instances.length > 1) fail('thread-instance-ambiguous');", + "new": "if (false) fail('thread-instance-ambiguous');", + "test": "packages/bus/tests/threads.test.mjs", + "pattern": "C2-09a: explicit creative" + }, + { + "id": "T05", + "file": "packages/bus/src/broker.mjs", + "old": "!this.#granted(s, 'thread.create', identifier(a.project), instance)", + "new": "!this.#granted(s, 'thread.create', identifier(a.project), null)", + "test": "packages/bus/tests/user-principal.test.mjs", + "pattern": "C2-09a: naming an instance" + }, + { + "id": "T06", + "file": "packages/cli/src/launcher.mjs", + "old": "const threaded = THREAD_DEFINITIONS.includes(b.roles[instance]?.definition);", + "new": "const threaded = b.roles[instance]?.definition === \"assistant\";", + "test": "packages/cli/tests/thread-launch.test.mjs", + "pattern": "C2-09a: creative thread messages" + }, + { + "id": "T07", + "file": "packages/webui/src/serve.mjs", + "old": ".filter(([, entry]) => THREAD_DEFINITIONS.includes(entry?.definition))", + "new": ".filter(() => true)", + "test": "packages/webui/tests/chat-instances.test.mjs", + "pattern": "C2-09a metadata uses the human" + }, + { + "id": "T08", + "file": "packages/webui/src/serve.mjs", + "old": "name: entry.vars?.['persona.name'] || instance", + "new": "name: instance", + "test": "packages/webui/tests/chat-instances.test.mjs", + "pattern": "C2-09a metadata uses the human" + }, + { + "id": "T09", + "file": "packages/webui/src/chat-http.mjs", + "old": "(Object.hasOwn(v, 'instance') && (typeof v.instance !== 'string' || !KEY.test(v.instance)))", + "new": "false", + "test": "packages/webui/tests/chat.test.mjs", + "pattern": "C2-09a malformed instance" + }, + { + "id": "T10", + "file": "packages/webui/src/chat.mjs", + "old": "createThread: ({ project, name, instance }) => channel('thread.create', { project, name, ...(instance === undefined ? {} : { instance }) }),", + "new": "createThread: ({ project, name, instance }) => channel('thread.create', { project, name }),", + "test": "packages/webui/tests/chat.test.mjs", + "pattern": "C2-09a createThread forwards" + }, + { + "id": "T11", + "file": "packages/webui/src/public/chat.js", + "old": ": { project: key, name: f.name, instance: f.instance };", + "new": ": { project: key, name: f.name };", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: no silent choice" + }, + { + "id": "T12", + "file": "packages/webui/src/public/chat.js", + "old": "!f.instance && listed.rows.length === 1", + "new": "!f.instance && listed.rows.length >= 1", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: no silent choice" + }, + { + "id": "T13", + "file": "packages/webui/src/public/chat.js", + "old": "current.instanceError = \"Couldn't read the thread instances; reload to try again\";", + "new": "current.instanceError = '';", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: metadata failure" + }, + { + "id": "T14", + "file": "packages/webui/src/public/chat.js", + "old": "txt(row.name === row.instance ? row.instance : `${row.name} (${row.instance})`)", + "new": "txt(row.name)", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: no silent choice" + }, + { + "id": "T15", + "file": "packages/webui/src/public/chat.js", + "old": "if (f.instance && !listed.rows.some(row => row.instance === f.instance)) {", + "new": "if (false) {", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: metadata failure" + }, + { + "id": "T16", + "file": "packages/webui/src/public/chat.js", + "old": "f.instance = ''; f.instanceChosen = true;", + "new": "f.instance = ''; f.instanceChosen = false;", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: metadata failure" + }, + { + "id": "T17", + "file": "packages/webui/src/public/chat.js", + "old": "const unavailable = !project && (!!current.instanceError || !choices.some(row => row.instance === f.instance));", + "new": "const unavailable = false;", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: no silent choice" + }, + { + "id": "T18", + "file": "packages/webui/src/serve.mjs", + "old": "if (req.method !== 'GET') return json(res, 405, { error: 'method-not-allowed' });", + "new": "if (false) return json(res, 405, { error: 'method-not-allowed' });", + "test": "packages/webui/tests/chat-instances.test.mjs", + "pattern": "C2-09a metadata uses the human" + }, + { + "id": "T19", + "file": "packages/webui/src/serve.mjs", + "old": "if (search) return json(res, 400, { error: 'invalid-request' });", + "new": "if (false) return json(res, 400, { error: 'invalid-request' });", + "test": "packages/webui/tests/chat-instances.test.mjs", + "pattern": "C2-09a metadata uses the human" + }, + { + "id": "T20", + "file": "packages/webui/src/public/chat.js", + "old": "current.instances = listed.rows; current.instanceError = '';", + "new": "current.instances = listed.rows;", + "test": "packages/webui/tests/chat-browser.test.mjs", + "pattern": "C2-09a picker: metadata failure" + }, + { + "id": "T21", + "file": "packages/webui/src/serve.mjs", + "old": "({ instance, name: entry.vars?.['persona.name'] || instance })", + "new": "({ instance, name: entry.vars?.['persona.name'] || instance, vars: entry.vars })", + "test": "packages/webui/tests/chat-instances.test.mjs", + "pattern": "C2-09a metadata uses the human" + }, + { + "id": "T22", + "file": "packages/cli/src/launcher.mjs", + "old": "const threaded = THREAD_DEFINITIONS.includes(b.roles[r.instance]?.definition);", + "new": "const threaded = b.roles[r.instance]?.definition === \"assistant\";", + "test": "packages/cli/tests/thread-launch.test.mjs", + "pattern": "C2-09a: terminating one creative thread" + }, + { + "id": "T23", + "file": "packages/cli/src/launcher.mjs", + "old": "const threaded = THREAD_DEFINITIONS.includes(b.roles[instance]?.definition);", + "new": "const threaded = b.roles[instance]?.definition === \"assistant\";", + "test": "packages/cli/tests/thread-launch.test.mjs", + "pattern": "C2-09a: fromSession refuses a creative instance outside a thread without a sandbox", + "additional_edits": [ + { + "file": "packages/cli/src/launcher.mjs", + "old": "const threaded = THREAD_DEFINITIONS.includes(b.roles[r.instance]?.definition);", + "new": "const threaded = b.roles[r.instance]?.definition === \"assistant\";" + } + ] + } +] diff --git a/agents/rocko/work/queue-219/old-persona-test.diff b/agents/rocko/work/queue-219/old-persona-test.diff new file mode 100644 index 000000000..84128f45a --- /dev/null +++ b/agents/rocko/work/queue-219/old-persona-test.diff @@ -0,0 +1,40 @@ +diff --git a/packages/cli/tests/launcher.test.mjs b/packages/cli/tests/launcher.test.mjs +index 2202c7753..0a34fe27a 100644 +--- a/packages/cli/tests/launcher.test.mjs ++++ b/packages/cli/tests/launcher.test.mjs +@@ -773,21 +773,28 @@ test("row 213 a creative instance's persona is read from /personas: o + const dir = join(given.root, "config", "personas", "acme"); + mkdirSync(dir, { recursive: true, mode: 0o700 }); + writeFileSync(join(dir, "bean.md"), "Not the pinned text.\n", { mode: 0o600 }); +- const { f, host, launcher } = await setup(t, undefined, given); +- // The chief of staff holds role.launch; a test chief session bound to this process asks. +- const chief = await host.bindLaunch({ business: "acme", role: "chief", run: "chief-test", harness: "pi", pid: process.pid, startTime: startTimeOf(process.pid) }); +- await new Client({ path: host.path, cap: chief.cap }).call("role.claim"); +- // The bundle's refusal comes back as launch-failed, as webPlan's does, with its reason in the log. +- assert.deepEqual(await launcher.fromSession(chief.cap, "bean"), { ok: false, error: "launch-failed" }); ++ // Row 219: creative instances launch from their thread, not from a chief's standalone request. ++ // The mismatch occurs before bindLaunch, so only the host's thread-state snapshot is a fixture. ++ const thread = 'persona-thread'; ++ const { f, launcher } = await setup(t, undefined, given, { wrap: host => ({ ++ ...host, ++ op: message => message.op === 'threadsState' ++ ? Promise.resolve([{ thread, instance: 'bean', holder: null, pending: 1 }]) ++ : host.op(message), ++ }) }); ++ // The bundle's refusal remains launch-failed with the persona mismatch as its reason. ++ await launcher.tick(); + const [e] = log(f); + assert.deepEqual([e.event, e.code, log(f).length], ["refused", "launch-failed", 1]); ++ assert.equal(e.thread, thread); + assert.equal(e.why, `instance bean: persona file ${join(dir, "bean.md")} isn't the file persona.sha256 names`); + assert.ok(!JSON.stringify(log(f)).includes("PERSONA-MARKER-d2")); + assert.deepEqual(readSessions(f.dataRoot), []); + // The run directory exists by then (BUILD.md residual); the persona is read before any bundle file. + const runs = readdirSync(join(f.dataRoot, "launches", "acme")); + assert.equal(runs.length, 1); +- assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), ["session"]); ++ assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), []); ++ assert.equal(existsSync(join(f.dataRoot, "threads", "acme", thread, "session")), true); + }); + + test("row 64 on Claude Code a role skill that isn't enabled refuses the same way, before anything starts", { timeout: 30000 }, async (t) => { diff --git a/agents/rocko/work/queue-219/prepare-mutants.py b/agents/rocko/work/queue-219/prepare-mutants.py new file mode 100644 index 000000000..63263d48f --- /dev/null +++ b/agents/rocko/work/queue-219/prepare-mutants.py @@ -0,0 +1,39 @@ +"""Static preparation only. Executing tests requires Sage CLEAR.""" +from pathlib import Path +import json +root=Path(__file__).resolve().parents[4] +p=Path(__file__).resolve().parent +specs=[] +def site(id,file,old,new,test,pattern): + text=(root/file).read_text();assert text.count(old)==1,(id,text.count(old)) + specs.append(dict(id=id,file=file,old=old,new=new,test=test,pattern=pattern)) +bus='packages/bus/src/broker.mjs';bt='packages/bus/tests/threads.test.mjs';cli='packages/cli/src/launcher.mjs';ct='packages/cli/tests/thread-launch.test.mjs';web='packages/webui/src/serve.mjs';wt='packages/webui/tests/chat-instances.test.mjs';browser='packages/webui/src/public/chat.js';brtest='packages/webui/tests/chat-browser.test.mjs' +site('T01',bus,'return THREAD_DEFINITIONS.includes(this.#business(business).roles[role]?.definition);',"return this.#business(business).roles[role]?.definition === 'assistant';",bt,'C2-09a: explicit creative') +site('T02',bus,': identifier(a.instance);',': instances[0];',bt,'C2-09a: explicit creative') +site('T03',bus,"if (!instances.length || (a.instance !== undefined && !instances.includes(instance))) fail('thread-instance-missing');","if (!instances.length) fail('thread-instance-missing');",bt,'C2-09a: explicit missing') +site('T04',bus,"if (a.instance === undefined && instances.length > 1) fail('thread-instance-ambiguous');","if (false) fail('thread-instance-ambiguous');",bt,'C2-09a: explicit creative') +site('T05',bus,"!this.#granted(s, 'thread.create', identifier(a.project), instance)","!this.#granted(s, 'thread.create', identifier(a.project), null)",'packages/bus/tests/user-principal.test.mjs','C2-09a: naming an instance') +site('T06',cli,'const threaded = THREAD_DEFINITIONS.includes(b.roles[instance]?.definition);','const threaded = b.roles[instance]?.definition === "assistant";',ct,'C2-09a: creative thread messages') +site('T07',web,'.filter(([, entry]) => THREAD_DEFINITIONS.includes(entry?.definition))','.filter(() => true)',wt,'C2-09a metadata uses the human') +site('T08',web,"name: entry.vars?.['persona.name'] || instance",'name: instance',wt,'C2-09a metadata uses the human') +site('T09','packages/webui/src/chat-http.mjs',"(Object.hasOwn(v, 'instance') && (typeof v.instance !== 'string' || !KEY.test(v.instance)))",'false','packages/webui/tests/chat.test.mjs','C2-09a malformed instance') +site('T10','packages/webui/src/chat.mjs',"createThread: ({ project, name, instance }) => channel('thread.create', { project, name, ...(instance === undefined ? {} : { instance }) }),","createThread: ({ project, name, instance }) => channel('thread.create', { project, name }),",'packages/webui/tests/chat.test.mjs','C2-09a createThread forwards') +site('T11',browser,": { project: key, name: f.name, instance: f.instance };",": { project: key, name: f.name };",brtest,'C2-09a picker: no silent choice') +site('T12',browser,'!f.instance && listed.rows.length === 1','!f.instance && listed.rows.length >= 1',brtest,'C2-09a picker: no silent choice') +site('T13',browser,'current.instanceError = "Couldn\'t read the thread instances; reload to try again";',"current.instanceError = '';",brtest,'C2-09a picker: metadata failure') +site('T14',browser,'txt(row.name === row.instance ? row.instance : `${row.name} (${row.instance})`)','txt(row.name)',brtest,'C2-09a picker: no silent choice') +site('T15',browser,'if (f.instance && !listed.rows.some(row => row.instance === f.instance)) {','if (false) {',brtest,'C2-09a picker: metadata failure') +site('T16',browser,"f.instance = ''; f.instanceChosen = true;","f.instance = ''; f.instanceChosen = false;",brtest,'C2-09a picker: metadata failure') +site('T17',browser,"const unavailable = !project && (!!current.instanceError || !choices.some(row => row.instance === f.instance));",'const unavailable = false;',brtest,'C2-09a picker: no silent choice') +# Route-local guards include their full unique response text to avoid touching other routes. +site('T18',web,"if (req.method !== 'GET') return json(res, 405, { error: 'method-not-allowed' });","if (false) return json(res, 405, { error: 'method-not-allowed' });",wt,'C2-09a metadata uses the human') +site('T19',web,"if (search) return json(res, 400, { error: 'invalid-request' });","if (false) return json(res, 400, { error: 'invalid-request' });",wt,'C2-09a metadata uses the human') +site('T20',browser,"current.instances = listed.rows; current.instanceError = '';","current.instances = listed.rows;",brtest,'C2-09a picker: metadata failure') +site('T21',web,"({ instance, name: entry.vars?.['persona.name'] || instance })","({ instance, name: entry.vars?.['persona.name'] || instance, vars: entry.vars })",wt,'C2-09a metadata uses the human') +site('T22',cli,'const threaded = THREAD_DEFINITIONS.includes(b.roles[r.instance]?.definition);','const threaded = b.roles[r.instance]?.definition === "assistant";',ct,'C2-09a: terminating one creative thread') +site('T23',cli,'const threaded = THREAD_DEFINITIONS.includes(b.roles[instance]?.definition);','const threaded = b.roles[instance]?.definition === "assistant";',ct,'C2-09a: fromSession refuses a creative instance outside a thread without a sandbox') +ending_old='const threaded = THREAD_DEFINITIONS.includes(b.roles[r.instance]?.definition);' +assert (root/cli).read_text().count(ending_old)==1 +specs[-1]['additional_edits']=[dict(file=cli,old=ending_old,new='const threaded = b.roles[r.instance]?.definition === "assistant";')] +(p/'mutants.json').write_text(json.dumps(specs,indent=2)+'\n') +print('STATIC_PREFLIGHT',len(specs),'exact-once sites; no tests run') diff --git a/packages/bus/src/broker.mjs b/packages/bus/src/broker.mjs index c50f8ac23..5ee242f1d 100644 --- a/packages/bus/src/broker.mjs +++ b/packages/bus/src/broker.mjs @@ -4,7 +4,7 @@ import { initializeControl, processingStatus } from '../../processing-control/sr import { NAME as HARNESS_NAME } from '../../plugins/src/kinds.mjs'; import { loadRequirements } from '../../tasks/src/requirements.mjs'; import { MERGE, LandingRefusal, readLanding } from './landing.mjs'; -import { ID_PATTERN, THREAD_DEFINITION, END_PROTECTED, forkProblem, forkUrl, credentialKind, validArgv, commandAllowed, HOST_ID, CREATIVE_DEFINITIONS, unhireableAction, giteaName, parseGiteaUrl, covers } from '../../business/src/vocabulary.mjs'; +import { ID_PATTERN, THREAD_DEFINITIONS, END_PROTECTED, forkProblem, forkUrl, credentialKind, validArgv, commandAllowed, HOST_ID, CREATIVE_DEFINITIONS, unhireableAction, giteaName, parseGiteaUrl, covers } from '../../business/src/vocabulary.mjs'; // Kept closed while S1 lands; its vocabulary is the integration authority. export const ACTIONS = Object.freeze([ @@ -312,7 +312,7 @@ const enrollBinding = (enroll) => ({ }); // r6 E3: the installer line's comment, then the tag's 32 hex digits. const INSTALLER_COMMENT = 'mosaic-installer-'; -// Row 121 (#1615): an instance of THREAD_DEFINITION (vocabulary.mjs since row 155) talks with the +// Row 121 (#1615): an instance of THREAD_DEFINITIONS (vocabulary.mjs since row 213) talks with the // user in threads. Its runs claim a thread, not the role, and every message to or from it names its // thread (REQ-MSG-1). // Row 146 (#1640): the sender of the bus's notice to launch.by. A role instance name matches the @@ -593,7 +593,7 @@ export class Broker { ); } #threaded(business, role) { - return this.#business(business).roles[role]?.definition === THREAD_DEFINITION; + return THREAD_DEFINITIONS.includes(this.#business(business).roles[role]?.definition); } // Row 264a (#1769): a claim holds one session of a thread, and each session has its own holder. #sessionHolder(session) { @@ -5790,24 +5790,27 @@ export class Broker { .all('SELECT id,at,name,repo FROM projects WHERE business=? ORDER BY seq', s.business) .filter((p) => !s.user || this.#granted(s, 'project.list', p.id, null)); } - // A thread belongs to the user who made it and to the business's one thread instance. + // A thread belongs to the user who made it and to one declared thread instance. // Row 177 (#1675): a tester's thread is owned by `tester`, so neither reaches the other's. case 'thread.create': { if (!s.tester && !s.user) this.#human(s); - keys(a, ['project', 'name'], ['project', 'name']); + keys(a, ['project', 'name', 'instance'], ['project', 'name']); if (!name(a.name)) fail('invalid-request'); + const instances = Object.keys(b.roles).filter((r) => this.#threaded(s.business, r)); + const instance = a.instance === undefined + ? (instances.length === 1 ? instances[0] : null) + : identifier(a.instance); // FED R3 (design 3.5): a user's grant must cover the project and the thread instance, checked // before the project is looked up, so an ungranted id reads the same whether it exists or not. if (s.user) { - const threaded = Object.keys(b.roles).filter((r) => this.#threaded(s.business, r)); - if (threaded.length !== 1 || !this.#granted(s, 'thread.create', identifier(a.project), threaded[0])) fail('access-refused'); + if (!instance || !instances.includes(instance) || !this.#granted(s, 'thread.create', identifier(a.project), instance)) + fail('access-refused'); } if (!this.#store.get('SELECT 1 FROM projects WHERE id=? AND business=?', identifier(a.project), s.business)) fail('project-not-found'); - // The bus never picks between instances; the business file declares exactly one. - const instances = Object.keys(b.roles).filter((r) => this.#threaded(s.business, r)); - if (!instances.length) fail('thread-instance-missing'); - if (instances.length > 1) fail('thread-instance-ambiguous'); + // The human selects an instance; omission is compatible only with one candidate. + if (!instances.length || (a.instance !== undefined && !instances.includes(instance))) fail('thread-instance-missing'); + if (a.instance === undefined && instances.length > 1) fail('thread-instance-ambiguous'); if (this.#store.get('SELECT 1 FROM threads WHERE project=? AND name=?', a.project, a.name)) fail('thread-exists'); const tid = randomUUID(); this.#store.run( @@ -5816,13 +5819,13 @@ export class Broker { this.#now(), s.business, a.project, - instances[0], + instance, this.#owner(s), a.name, ); // Row 264a (#1769, D3): every thread has a default session from the start. - this.#newSession(s, { id: tid, business: s.business, instance: instances[0], owner: this.#owner(s) }, 'default', 'default', true); - return { id: tid, instance: instances[0] }; + this.#newSession(s, { id: tid, business: s.business, instance, owner: this.#owner(s) }, 'default', 'default', true); + return { id: tid, instance }; } // Row 177 (#1675): each lists its own threads. The human's list leaves out the tester's, // and the tester's lists only its own (row 175 listed it the business human's). diff --git a/packages/bus/tests/creative.test.mjs b/packages/bus/tests/creative.test.mjs index 6cfbf1cd6..6d421ba19 100644 --- a/packages/bus/tests/creative.test.mjs +++ b/packages/bus/tests/creative.test.mjs @@ -44,15 +44,18 @@ function setup(t) { const caps = {}; caps.human = b.bindHuman({ business: 'demo', human: 'jason', via: 'cli', outsideAgent: true }); const project = b.request(caps.human, { verb: 'project.create', args: { name: 'demo' } }).id; - const thread = b.request(caps.human, { verb: 'thread.create', args: { project, name: 'main' } }).id; + const thread = b.request(caps.human, { verb: 'thread.create', args: { project, name: 'main', instance: 'assistant' } }).id; + const threads = { assistant: thread }; for (const role of Object.keys(roles)) { - caps[role] = b.bindLaunch({ business: 'demo', role, run: `${role}-run`, harness: 'pi', ...(role === 'assistant' ? { thread } : {}) }); + if (['bean', 'quill'].includes(role)) + threads[role] = b.request(caps.human, { verb: 'thread.create', args: { project, name: role, instance: role } }).id; + caps[role] = b.bindLaunch({ business: 'demo', role, run: `${role}-run`, harness: 'pi', ...(threads[role] ? { thread: threads[role] } : {}) }); b.request(caps[role], { verb: 'role.claim' }); } caps.tester = b.bindTester({ business: 'demo' }); const call = (who, verb, args = {}) => b.request(caps[who], { verb, args }); const count = (sql) => store.get(sql).n; - return { call, count, store }; + return { call, count, store, threads }; } const raise = { action: 'policy.change', question: 'Change?', options: [{ key: 'yes', text: 'Yes' }, { key: 'no', text: 'No' }], recommendation: 'yes', blocking: false }; @@ -72,7 +75,7 @@ test('M6: a creative raise refuses decision-not-allowed, cited or not, and no de }); test('a creative instance sends only to the user and hears only from the user', (t) => { - const { call, count } = setup(t); + const { call, count, threads } = setup(t); for (const who of ['bean', 'quill']) { call(who, 'message.send', { to: 'human', body: 'hello' }); for (const to of ['chief', 'assistant', 'pm', 'bean', 'quill'].filter((n) => n !== who)) { @@ -81,7 +84,7 @@ test('a creative instance sends only to the user and hears only from the user', for (const from of ['chief', 'pm']) { assert.throws(() => call(from, 'message.send', { to: who, body: 'hello' }), { code: 'message-sender-refused' }, `${from} to ${who}`); } - call('human', 'message.send', { to: who, body: 'hello' }); + call('human', 'thread.send', { thread: threads[who], body: 'hello' }); } assert.equal(count("SELECT count(*) n FROM messages WHERE from_role IN ('bean','quill')"), 2); assert.equal(count("SELECT count(*) n FROM messages WHERE to_role IN ('bean','quill')"), 2); diff --git a/packages/bus/tests/threads.test.mjs b/packages/bus/tests/threads.test.mjs index 8b31da9f9..1410ec3fa 100644 --- a/packages/bus/tests/threads.test.mjs +++ b/packages/bus/tests/threads.test.mjs @@ -601,3 +601,42 @@ test('project.create keeps unrelated names with and without a repository', t => { id: linked.id, name: 'work', repo: 'stack' }, ]); }); + +test('C2-09a: explicit creative instances bind independent threads, default sessions and holders', t => { + const creative = ['companion', 'ideas', 'research', 'proofreader', 'fact-checker', 'continuity']; + const extra = Object.fromEntries(creative.map((definition, i) => [`writer${i}`, { definition, authority: { withinRole: ['message.send'], crossRole: [] } }])); + const ctx = setup(t, demo({ roles: { ...roles, ...extra }, definitions: { + ...definitions, ...Object.fromEntries(creative.map(definition => [definition, { messaging: shipped(definition).messaging }])), + } })); + const project = ctx.call('human', 'project.create', { name: 'Writing' }).id; + assert.throws(() => ctx.call('human', 'thread.create', { project, name: 'ambiguous' }), { code: 'thread-instance-ambiguous' }); + for (const [instance, role] of Object.entries(extra)) { + const receipt = ctx.call('human', 'thread.create', { project, name: instance, instance }); + assert.equal(receipt.instance, instance); + assert.equal(ctx.store.get('SELECT instance FROM threads WHERE id=?', receipt.id).instance, instance); + assert.equal(ctx.store.get('SELECT instance FROM sessions WHERE thread=?', receipt.id).instance, instance); + ctx.launch(instance, instance, { thread: receipt.id }); + assert.equal(ctx.claim(instance).thread, receipt.id); + const sent = ctx.call('human', 'thread.send', { thread: receipt.id, body: `For ${instance}` }); + const read = ctx.call(instance, 'message.receive'); + assert.deepEqual(read.map(m => m.id), [sent.id], role.definition); + assert.equal(ctx.threadOf(sent.id), receipt.id); + assert.equal(ctx.store.get('SELECT count(*) n FROM role_claims WHERE role=?', instance).n, 0); + assert.throws(() => ctx.call('human', 'message.send', { to: instance, body: 'No implicit thread' }), + { code: 'thread-required-use-thread-send' }); + } + assert.equal(ctx.count('SELECT count(*) n FROM threads'), creative.length); +}); + +test('C2-09a: explicit missing, non-thread and malformed selectors create no thread or session', t => { + const ctx = setup(t); + const project = ctx.call('human', 'project.create', { name: 'Writing' }).id; + for (const instance of ['not-declared', 'pm', 'coder']) + assert.throws(() => ctx.call('human', 'thread.create', { project, name: instance, instance }), { code: 'thread-instance-missing' }); + for (const instance of [null, 7, {}, [], '', 'bad name']) + assert.throws(() => ctx.call('human', 'thread.create', { project, name: 'bad', instance }), { code: 'invalid-request' }); + assert.equal(ctx.count('SELECT count(*) n FROM threads'), 0); + assert.equal(ctx.count('SELECT count(*) n FROM sessions'), 0); + const legacy = ctx.call('human', 'thread.create', { project, name: 'Default' }); + assert.equal(legacy.instance, 'assistant'); +}); diff --git a/packages/bus/tests/user-principal.test.mjs b/packages/bus/tests/user-principal.test.mjs index 4dcd1ebeb..8471ddf85 100644 --- a/packages/bus/tests/user-principal.test.mjs +++ b/packages/bus/tests/user-principal.test.mjs @@ -23,7 +23,7 @@ const grant = (over = {}) => ({ }); // `processing` stands in for the halt switch, as lifecycle-verbs.test.mjs's T15 does. -function setup(t, { grants = [grant()], processing = null } = {}) { +function setup(t, { grants = [grant()], processing = null, threadRoles = {} } = {}) { const root = mkdtempSync(join(tmpdir(), 'bus-user-principal-')); const store = new Store(root); t.after(() => { @@ -40,6 +40,7 @@ function setup(t, { grants = [grant()], processing = null } = {}) { roles: { chief: { definition: 'chief-of-staff', authority: { withinRole: ['message.send'], crossRole: [] } }, assistant: { definition: 'assistant', authority: { withinRole: ['message.send'], crossRole: [] } }, + ...threadRoles, }, definitions: { 'chief-of-staff': {}, assistant: { messaging: shipped('assistant').messaging } }, }, @@ -394,3 +395,21 @@ test('R3 r4: the chief\'s hire refuses an instance named user with the code it g test('R3 r4: a non-function access throws before the store is read', () => { assert.throws(() => new Broker({ store: null, businesses: { demo: {} }, access: 'access.json' }), /access must be a function/); }); + +test('C2-09a: naming an instance cannot bypass its collaborator grant or disclose an ungranted project', t => { + const ctx = setup(t, { threadRoles: { bean: { definition: 'companion', authority: { withinRole: ['message.send'], crossRole: [] } } } }); + const project = ctx.call('human', 'project.create', { name: 'shared' }).id; + const privateProject = ctx.call('human', 'project.create', { name: 'private' }).id; + ctx.access.doc.grants = [grant({ projects: [project] })]; + ctx.user('alice'); + assert.equal(ctx.code('alice', 'thread.create', { project, name: 'blocked bean', instance: 'bean' }), 'access-refused'); + for (const id of [privateProject, '00000000-0000-4000-8000-000000000000']) + assert.equal(ctx.code('alice', 'thread.create', { project: id, name: 'blocked project', instance: 'assistant' }), 'access-refused'); + assert.equal(ctx.store.get('SELECT count(*) n FROM threads').n, 0); + const allowed = ctx.call('alice', 'thread.create', { project, name: 'assistant work', instance: 'assistant' }); + assert.equal(allowed.instance, 'assistant'); + ctx.access.doc.grants = [grant({ projects: [project], instances: ['bean'] })]; + const bean = ctx.call('alice', 'thread.create', { project, name: 'bean work', instance: 'bean' }); + assert.equal(bean.instance, 'bean'); + assert.equal(ctx.code('alice', 'thread.create', { project, name: 'old instance', instance: 'assistant' }), 'access-refused'); +}); diff --git a/packages/cli/src/launcher.mjs b/packages/cli/src/launcher.mjs index e044b7627..f019df83f 100644 --- a/packages/cli/src/launcher.mjs +++ b/packages/cli/src/launcher.mjs @@ -53,7 +53,7 @@ // session.json names it as gitSocket. // // Thread sessions (row 121, #1615). A thread instance (an instance whose -// role definition is the assistant) runs once per thread, never on its own: +// role definition is in THREAD_DEFINITIONS) runs once per thread, never on its own: // a launch from a session or --pm refuses thread-required. Every tick // (threadTickMs) the host asks the broker for the business's threads // (launch op threadsState) and: @@ -139,6 +139,7 @@ import { createServer } from "node:net"; import { chmodSync, existsSync, lstatSync, mkdirSync, readFileSync, renameSync, rmSync } from "node:fs"; import { dirname, join, relative } from "node:path"; import { configDir, loadBusiness, resolveInstance, systemVars } from "../../business/src/index.mjs"; +import { THREAD_DEFINITIONS } from "../../business/src/vocabulary.mjs"; import { limitHit } from "../../business/src/launch.mjs"; import { BundleError, buildBundle, hasFallback, readSkills, sessionModel } from "../../harness/src/bundle.mjs"; import { createLoader } from "../../plugins/src/loader.mjs"; @@ -175,8 +176,6 @@ const STOP_TIMEOUT_MS = 30000; const LINE_MAX = 4096; // How often the host asks for the threads' state; a poll period, not a limit. const THREAD_TICK_MS = 2000; -// The broker's thread instance definition (packages/bus/src/broker.mjs). -const THREAD_DEFINITION = "assistant"; // Row 194: the broker's LIFECYCLE_REQUEST_MS. A regenerate's old run must be gone by then. const LIFECYCLE_REQUEST_MS = 60000; @@ -299,7 +298,7 @@ export function createLauncher({ try { b=loadBusiness(host.business,{dir:configDir(env),rolesDir:rolesDir(env)}); } catch(e){throw new Refusal("organization-refused",e.message);} if (!b.launch) throw new Refusal("launch-not-configured", `business ${b.id} has no launch block`); if (!(pm ? instance === b.launch.by : b.launch.instances.includes(instance) || (human && Object.hasOwn(b.roles, instance)))) throw new Refusal("instance-not-listed"); - const threaded = b.roles[instance]?.definition === THREAD_DEFINITION; + const threaded = THREAD_DEFINITIONS.includes(b.roles[instance]?.definition); if (threaded && !thread) throw new Refusal("thread-required", `${instance} runs only for a thread`); if (!threaded && thread) throw new Refusal("thread-not-allowed", `${instance} isn't a thread instance`); if (threaded && !session) throw new Refusal("session-required", `${instance} runs for one session of a thread`); @@ -854,7 +853,7 @@ export function createLauncher({ function ending(r) { const runs = r.runs ?? []; const at = Date.parse(r.at); - const threaded = b.roles[r.instance]?.definition === THREAD_DEFINITION; + const threaded = THREAD_DEFINITIONS.includes(b.roles[r.instance]?.definition); return registry.sessions.filter((s) => s.business === b.id && (runs.includes(s.run) || (!threaded && s.instance === r.instance && (spawnedAt.get(s.run) ?? Date.parse(s.startedAt)) <= at))); } diff --git a/packages/cli/tests/launcher.test.mjs b/packages/cli/tests/launcher.test.mjs index 2202c7753..0a34fe27a 100644 --- a/packages/cli/tests/launcher.test.mjs +++ b/packages/cli/tests/launcher.test.mjs @@ -773,21 +773,28 @@ test("row 213 a creative instance's persona is read from /personas: o const dir = join(given.root, "config", "personas", "acme"); mkdirSync(dir, { recursive: true, mode: 0o700 }); writeFileSync(join(dir, "bean.md"), "Not the pinned text.\n", { mode: 0o600 }); - const { f, host, launcher } = await setup(t, undefined, given); - // The chief of staff holds role.launch; a test chief session bound to this process asks. - const chief = await host.bindLaunch({ business: "acme", role: "chief", run: "chief-test", harness: "pi", pid: process.pid, startTime: startTimeOf(process.pid) }); - await new Client({ path: host.path, cap: chief.cap }).call("role.claim"); - // The bundle's refusal comes back as launch-failed, as webPlan's does, with its reason in the log. - assert.deepEqual(await launcher.fromSession(chief.cap, "bean"), { ok: false, error: "launch-failed" }); + // Row 219: creative instances launch from their thread, not from a chief's standalone request. + // The mismatch occurs before bindLaunch, so only the host's thread-state snapshot is a fixture. + const thread = 'persona-thread'; + const { f, launcher } = await setup(t, undefined, given, { wrap: host => ({ + ...host, + op: message => message.op === 'threadsState' + ? Promise.resolve([{ thread, instance: 'bean', holder: null, pending: 1 }]) + : host.op(message), + }) }); + // The bundle's refusal remains launch-failed with the persona mismatch as its reason. + await launcher.tick(); const [e] = log(f); assert.deepEqual([e.event, e.code, log(f).length], ["refused", "launch-failed", 1]); + assert.equal(e.thread, thread); assert.equal(e.why, `instance bean: persona file ${join(dir, "bean.md")} isn't the file persona.sha256 names`); assert.ok(!JSON.stringify(log(f)).includes("PERSONA-MARKER-d2")); assert.deepEqual(readSessions(f.dataRoot), []); // The run directory exists by then (BUILD.md residual); the persona is read before any bundle file. const runs = readdirSync(join(f.dataRoot, "launches", "acme")); assert.equal(runs.length, 1); - assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), ["session"]); + assert.deepEqual(readdirSync(join(f.dataRoot, "launches", "acme", runs[0])), []); + assert.equal(existsSync(join(f.dataRoot, "threads", "acme", thread, "session")), true); }); test("row 64 on Claude Code a role skill that isn't enabled refuses the same way, before anything starts", { timeout: 30000 }, async (t) => { diff --git a/packages/cli/tests/thread-launch.test.mjs b/packages/cli/tests/thread-launch.test.mjs index 0d207a0a0..84f8ed088 100644 --- a/packages/cli/tests/thread-launch.test.mjs +++ b/packages/cli/tests/thread-launch.test.mjs @@ -396,6 +396,22 @@ test("REQ-THREAD-1 REQ-LAUNCH-1 the chief can't launch the assistant outside a t assert.deepEqual(launchLog(f).map((e) => [e.event, e.instance, e.code]), [["refused", "assistant", "thread-required"]]); }); +// Admission refuses before bundle construction or a sandbox is needed. +test('C2-09a: fromSession refuses a creative instance outside a thread without a sandbox', { timeout: 30000 }, async t => { + const prepared = prepare(t, { edit: doc => { + doc.roles.bean = { ...doc.roles.assistant, definition: 'companion' }; + doc.launch.instances.push('bean'); + return doc; + } }); + const { f } = prepared; + const host = await start(t, prepared); + const chief = await host.host.bindLaunch({ business: 'acme', role: 'chief', run: 'chief-bean-refusal', harness: 'pi', pid: process.pid, startTime: startTimeOf(process.pid) }); + await new Client({ path: host.host.path, cap: chief.cap }).call('role.claim'); + assert.deepEqual(await host.launcher.fromSession(chief.cap, 'bean'), { ok: false, error: 'thread-required' }); + assert.deepEqual(readSessions(f.dataRoot), []); + assert.deepEqual(launchLog(f).map(e => [e.event, e.instance, e.code]), [['refused', 'bean', 'thread-required']]); +}); + // Row 125 (#1619, REQ-LIMIT-3, REQ-LIMIT-4): the token-limit check switched on, with the // assistant's limit at 0, so it is at its limit before any turn. The real check runs through the // bus's plug-in loader. @@ -443,3 +459,71 @@ test("REQ-LIMIT-4 a thread run at its token limit stops at the end of its turn: assert.deepEqual(threadRuns(f, A), []); assert.equal(host.runtime.threadsState("acme").find((x) => x.thread === A).holder, null); }); + +test('C2-09a: creative thread messages start distinct runs and keep each instance history separate', { skip, timeout: 90000 }, async t => { + const prepared = prepare(t, { sonnet: 2, idle: 60, edit: doc => { + const vars = { ...doc.roles.assistant.vars }; + delete doc.roles.assistant; + doc.roles.bean = { definition: 'companion', vars: { ...vars } }; + doc.roles.chet = { definition: 'continuity', vars: { ...vars } }; + doc.launch.instances = ['coder', 'bean', 'chet']; + return doc; + } }); + const host = await start(t, prepared), { f } = prepared; + const project = host.call('project.create', { name: 'Writing' }).id; + const bean = host.call('thread.create', { project, name: 'Bean', instance: 'bean' }).id; + const chet = host.call('thread.create', { project, name: 'Chet', instance: 'chet' }).id; + const send = async (thread, instance, body) => { + const message = host.call('thread.send', { thread, body }); + const outcome = await until(() => host.call('thread.messages', { thread }).find(m => m.in_reply_to === message.id) || + launchLog(f).find(row => row.thread === thread && row.event === 'refused')); + assert.notEqual(outcome.event, 'refused', `creative thread launch refused: ${outcome.code}`); + const reply = outcome; + assert.equal(reply.from_role, instance); + assert.equal(reply.to_role, 'human'); + return JSON.parse(reply.body.match(/THREAD (\{.*\})/)[1]); + }; + assert.deepEqual((await send(bean, 'bean', 'Bean input')).before, []); + assert.deepEqual((await send(chet, 'chet', 'Chet input')).before, []); + assert.deepEqual((await send(bean, 'bean', 'Bean again')).before, ['Bean input']); + assert.equal(readFileSync(join(f.dataRoot, 'threads', 'acme', bean, 'session', 'history'), 'utf8'), 'Bean input\nBean again\n'); + assert.equal(readFileSync(join(f.dataRoot, 'threads', 'acme', chet, 'session', 'history'), 'utf8'), 'Chet input\n'); + const launches = launchLog(f).filter(row => row.event === 'launch'); + assert.ok(launches.some(row => row.instance === 'bean' && row.thread === bean)); + assert.ok(launches.some(row => row.instance === 'chet' && row.thread === chet)); + const chief = await host.host.bindLaunch({ business: 'acme', role: 'chief', run: 'chief-creative-test', harness: 'pi', pid: process.pid, startTime: startTimeOf(process.pid) }); + await new Client({ path: host.host.path, cap: chief.cap }).call('role.claim'); + assert.deepEqual(await host.launcher.fromSession(chief.cap, 'bean'), { ok: false, error: 'thread-required' }); +}); + +test('C2-09a: terminating one creative thread does not signal an unselected peer of the same instance', { skip, timeout: 90000 }, async t => { + const prepared = prepare(t, { sonnet: 2, idle: 60, edit: doc => { + doc.roles.bean = { ...doc.roles.assistant, definition: 'companion' }; + delete doc.roles.assistant; + doc.launch.instances = ['coder', 'bean']; + return doc; + } }); + let requests = []; + const host = await start(t, prepared, { ops: () => ({ + lifecycleRequests: () => requests, + lifecycleTake: () => ({ ...requests[0], go: true }), + lifecycleDone: () => { requests = []; return {}; }, + }) }); + const { f } = prepared; + const project = host.call('project.create', { name: 'Writing' }).id; + const threads = ['one', 'two'].map(name => host.call('thread.create', { project, name, instance: 'bean' }).id); + for (const thread of threads) { + const message = host.call('thread.send', { thread, body: 'Stay in this thread' }); + const outcome = await until(() => host.call('thread.messages', { thread }).find(row => row.in_reply_to === message.id) || + launchLog(f).find(row => row.thread === thread && row.event === 'refused')); + assert.notEqual(outcome.event, 'refused', `creative thread launch refused: ${outcome.code}`); + } + const runs = threads.map(thread => launchLog(f).find(row => row.thread === thread && row.event === 'launch').run); + requests = [{ request: 'one-thread-end', state: 'open', business: 'acme', action: 'lifecycle.end', + instance: 'bean', runs: [runs[0]], by: f.doc.human, at: new Date(Date.now() + 1000).toISOString() }]; + await host.launcher.tick(); + await until(() => launchLog(f).some(row => row.event === 'lifecycle-end' && row.run === runs[0])); + assert.deepEqual(launchLog(f).filter(row => row.event === 'lifecycle-end').map(row => row.run), [runs[0]]); + await until(() => !readSessions(f.dataRoot).some(row => row.run === runs[0])); + assert.ok(readSessions(f.dataRoot).some(row => row.run === runs[1])); +}); diff --git a/packages/webui/src/chat-http.mjs b/packages/webui/src/chat-http.mjs index c7770726d..396b42f3c 100644 --- a/packages/webui/src/chat-http.mjs +++ b/packages/webui/src/chat-http.mjs @@ -78,7 +78,8 @@ export async function chatHTTP(req, res, { path, search, chat, named, send, prev (Object.hasOwn(v, 'repo') && (typeof v.repo !== 'string' || !KEY.test(v.repo)))) throw Error(); action = 'createProject'; args = v; } else if (path === '/api/chat/threads') { - if (!exact(v, ['project', 'name']) || !id(v.project) || !name(v.name)) throw Error(); + if (!exact(v, ['project', 'name', 'instance'], ['project', 'name']) || !id(v.project) || !name(v.name) || + (Object.hasOwn(v, 'instance') && (typeof v.instance !== 'string' || !KEY.test(v.instance)))) throw Error(); action = 'createThread'; args = v; } else if (match?.[2] === 'messages') { if (!exact(v, ['body']) || typeof v.body !== 'string' || !v.body.trim() || v.body.length > 32768) throw Error(); diff --git a/packages/webui/src/chat.mjs b/packages/webui/src/chat.mjs index 018574ad4..65b89dab3 100644 --- a/packages/webui/src/chat.mjs +++ b/packages/webui/src/chat.mjs @@ -20,7 +20,7 @@ export function createChat({ call }) { projects: () => channel('project.list', {}), createProject: ({ name, repo }) => channel('project.create', { name, ...(repo === undefined ? {} : { repo }) }), threads: ({ project } = {}) => channel('thread.list', project === undefined ? {} : { project }), - createThread: ({ project, name }) => channel('thread.create', { project, name }), + createThread: ({ project, name, instance }) => channel('thread.create', { project, name, ...(instance === undefined ? {} : { instance }) }), messages: ({ thread, after, limit }) => channel('thread.messages', { thread, ...(after ? { after } : {}), limit }), send: ({ thread, body }) => channel('thread.send', { thread, body, class: 'REQUEST' }), open: ({ thread }) => channel('thread.open', { thread }), diff --git a/packages/webui/src/public/chat.js b/packages/webui/src/public/chat.js index a6b805c1b..15066bdc6 100644 --- a/packages/webui/src/public/chat.js +++ b/packages/webui/src/public/chat.js @@ -14,7 +14,7 @@ return states.get(id); }; const form = key => { - if (!forms.has(key)) forms.set(key, { name: '', repo: '', receipt: '', pending: false }); + if (!forms.has(key)) forms.set(key, { name: '', repo: '', instance: '', instanceChosen: false, receipt: '', pending: false }); return forms.get(key); }; class ChatError extends Error { @@ -58,7 +58,11 @@ function creation(kind, key) { const f = form(key), project = kind === 'project'; const repo = project && current.repos.length ? `` : ''; - return `

New ${kind}

${repo}

${txt(f.receipt)}

`; + const choices = current.instances ?? []; + const picker = project ? '' : ``; + const unavailable = !project && (!!current.instanceError || !choices.some(row => row.instance === f.instance)); + const status = !project && current.instanceError ? current.instanceError : f.receipt || (!project && !choices.length ? 'No thread instances are configured.' : ''); + return `

New ${kind}

${repo}${picker}

${txt(status)}

`; } function threads(rows) { return rows.length ? `
    ${rows.map(t => `
  • ${link('threads', t.id, t.name)} ${txt(t.instance)}
    • `).join('')}
` : '

No threads yet.

'; @@ -106,7 +110,7 @@ const f = document.getElementById(keep.field.id); if (f && !f.disabled) { f.focus({ preventScroll: true }); - if (keep.field.start !== null) f.setSelectionRange(keep.field.start, keep.field.end); + if (keep.field.start != null && typeof f.setSelectionRange === 'function') f.setSelectionRange(keep.field.start, keep.field.end); } } else if (navigated) c.container.querySelector('h1')?.focus({ preventScroll: true }); } @@ -126,7 +130,7 @@ paint(); return; } const old = current; - current = { container, kind, id, canSend, projects: old?.projects ?? [], threads: old?.threads ?? [], repos: old?.repos ?? [], error: '', refused: false }; + current = { container, kind, id, canSend, projects: old?.projects ?? [], threads: old?.threads ?? [], repos: old?.repos ?? [], instances: old?.instances ?? [], instanceError: kind === 'project' ? 'Reading thread instances…' : '', error: '', refused: false }; if (navigated || !old) container.innerHTML = '

Reading chat…

'; try { const s = kind === 'thread' ? state(id) : null; @@ -147,6 +151,25 @@ s.receipt = 'Message shown in this conversation.'; s.receiptId = null; } } + if (kind === 'project') { + try { + const listed = await request('instances', undefined, signal); + if (g !== generation || out()) return; + if (!listed.rows.every(row => typeof row?.instance === 'string' && /^[a-zA-Z0-9][a-zA-Z0-9_.-]{0,159}$/.test(row.instance) && typeof row.name === 'string')) + throw new ChatError('invalid-response'); + current.instances = listed.rows; current.instanceError = ''; + const f = form(id); + if (f.instance && !listed.rows.some(row => row.instance === f.instance)) { + f.instance = ''; f.instanceChosen = true; + f.receipt = 'The selected instance is no longer available. Choose who this thread is with.'; + } + if (!f.instanceChosen && !f.instance && listed.rows.length === 1) f.instance = listed.rows[0].instance; + } catch (err) { + if (err.name === 'AbortError' || out() || g !== generation) return; + current.instances = []; + current.instanceError = "Couldn't read the thread instances; reload to try again"; + } + } // Repository choices are optional metadata; an unavailable read must not block unlinked chat. if (kind === 'list') { try { @@ -179,7 +202,11 @@ if (!current || out()) return; if (event.target.id === 'chat-body') state(current.id).draft = event.target.value; const f = event.target.closest('.chat-create'); - if (f && event.target.dataset.chatField) form(f.dataset.key)[event.target.dataset.chatField] = event.target.value; + if (f && event.target.dataset.chatField) { + const draft = form(f.dataset.key), field = event.target.dataset.chatField; + draft[field] = event.target.value; + if (field === 'instance') { draft.instanceChosen = true; draft.receipt = ''; paint(); } + } }); document.addEventListener('scroll', event => { if (event.target.matches?.('.chat-log') && current?.kind === 'thread') { @@ -232,7 +259,8 @@ void write('send', id, { body: state(id).draft }, `threads/${id}/messages`); } else if (event.target.classList.contains('chat-create')) { const key = event.target.dataset.key, f = form(key), project = key === 'projects'; - const value = project ? { name: f.name, ...(f.repo ? { repo: f.repo } : {}) } : { project: key, name: f.name }; + if (!project && (current.instanceError || !(current.instances ?? []).some(row => row.instance === f.instance))) return; + const value = project ? { name: f.name, ...(f.repo ? { repo: f.repo } : {}) } : { project: key, name: f.name, instance: f.instance }; void write(project ? 'createProject' : 'createThread', key, value, project ? 'projects' : 'threads'); } }); diff --git a/packages/webui/src/serve.mjs b/packages/webui/src/serve.mjs index 288871d0c..d9d27a062 100644 --- a/packages/webui/src/serve.mjs +++ b/packages/webui/src/serve.mjs @@ -8,6 +8,7 @@ import { BusReadError, busStatus as coreBusStatus, subjectOk, sendOutcome } from // Row 176, sage-1111: HTTP-only tester refusals; the row-164 transport pin stays untouched. const busStatus = code => ['tester-off', 'tester-required'].includes(code) ? 403 : coreBusStatus(code); import { ReadError, readStatus, fieldTable } from './reads.mjs'; +import { THREAD_DEFINITIONS } from '../../business/src/vocabulary.mjs'; import { chatHTTP } from './chat-http.mjs'; import { processingHTTP } from './processing-http.mjs'; @@ -705,6 +706,25 @@ export async function startServer({ host = '127.0.0.1', port = 7330, board = DEF if (req.method !== 'POST') return json(res, 405, { error: 'method not allowed' }); return busSend(req, res, { send, preview: signer, log }); } + // Row 219: labels only, from this listener's own human or tester bus, after auth above. + if (path === '/api/chat/instances') { + if (req.method !== 'GET') return json(res, 405, { error: 'method-not-allowed' }); + if (search) return json(res, 400, { error: 'invalid-request' }); + if (!bus?.instances) return json(res, 503, { error: 'not-configured' }); + try { + const value = await bus.instances(); + if (!value || !value.instances || typeof value.instances !== 'object' || Array.isArray(value.instances)) + throw new BusReadError('invalid-response'); + const rows = Object.entries(value.instances) + .filter(([, entry]) => THREAD_DEFINITIONS.includes(entry?.definition)) + .map(([instance, entry]) => ({ instance, name: entry.vars?.['persona.name'] || instance })); + if (rows.some(row => typeof row.name !== 'string')) throw new BusReadError('invalid-response'); + return json(res, 200, { rows, at: new Date().toISOString() }); + } catch (err) { + const code = err instanceof BusReadError ? err.code : 'invalid-response'; + return json(res, busStatus(code), { error: code }); + } + } if (path === '/api/chat' || path.startsWith('/api/chat/')) { return chatHTTP(req, res, { path, search, chat, named, send, preview: writer, log }); } diff --git a/packages/webui/tests/chat-browser.test.mjs b/packages/webui/tests/chat-browser.test.mjs index f3eeeca96..cc3ec093d 100644 --- a/packages/webui/tests/chat-browser.test.mjs +++ b/packages/webui/tests/chat-browser.test.mjs @@ -12,6 +12,7 @@ const HOSTILE = '\u001b[31m\u202eevil'; const stub = `(() => { const real = fetch.bind(window); window.__writes=[];window.__reads=[];window.__answers=[];window.__pages=[]; + window.__instances=[{instance:'assistant',name:'assistant'}]; window.__projects=[{id:'${P}', name:${JSON.stringify(HOSTILE)}, repo:null},{id:'\">',name:'Invalid record'}]; window.__threads=[{id:'${A}',project:'${P}',name:'alpha',instance:'assistant'}, {id:'${B}',project:'${P}',name:'beta',instance:'assistant'}]; window.fetch=async (url,init={})=>{ @@ -33,6 +34,7 @@ const stub = `(() => { window.__reads.push(path); if(window.__holdRead && path.includes(window.__holdRead)){window.__heldSignal=init.signal;await new Promise(r=>window.__releaseRead=r);} if(window.__readFailure)return new Response(JSON.stringify({error:window.__readFailure}),{status:window.__readFailure==='thread-not-owned'?403:503}); + if(path==='/api/chat/instances')return window.__instanceFailure ? new Response(JSON.stringify({error:'not-configured'}),{status:503}) : data({rows:window.__instances}); if(path==='/api/chat/projects')return data({rows:window.__projects}); if(path.startsWith('/api/chat/threads')&&!path.includes('/messages'))return data({rows:window.__threads}); if(window.__pages.length)return data(window.__pages.shift()); @@ -82,7 +84,7 @@ test('chat browser: list/create named projects and threads, hostile labels inert await b.evaluate(`window.__answers.push({body:{id:'${A}',instance:'assistant'}}),true`); await click('#chat-thread-form button[type=submit]'); await wait('document.querySelector("#chat-send-form")'); - assert.deepEqual(await b.evaluate('window.__writes[1].body'), { project: P, name: 'named thread' }); + assert.deepEqual(await b.evaluate('window.__writes[1].body'), { project: P, name: 'named thread', instance: 'assistant' }); assert.equal(await b.evaluate('window.__writes[1].path'), '/api/chat/threads'); assert.equal(await b.evaluate('window.__writes.length'), 2); }); @@ -264,3 +266,65 @@ test('chat browser: a late project creation receipt cannot navigate away from an assert.equal(await b.evaluate('location.hash'), '#/chat/threads/' + A); assert.equal(await b.evaluate('document.querySelector("#chat-body").value'), 'stay in alpha'); }); + +test('C2-09a picker: no silent choice, duplicate hostile labels stay inert, selection survives refresh and reaches POST', async t => { + const { b, wait, set, click, refresh } = await setup(t); + await b.evaluate(`window.__instances=[{instance:'bean',name:${JSON.stringify(HOSTILE)}},{instance:'chet',name:${JSON.stringify(HOSTILE)}}];location.hash='#/chat/projects/${P}';true`); + await wait('document.querySelectorAll("#chat-instance option").length===3'); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), ''); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").required'), true); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); + const labels = await b.evaluate('[...document.querySelectorAll("#chat-instance option")].map(o=>o.textContent)'); + assert.match(labels[1], /\(bean\)/); + assert.match(labels[2], /\(chet\)/); + assert.doesNotMatch(labels.join(''), /[\u202a-\u202e\u2066-\u2069]/u); + assert.equal(await b.evaluate('!!document.querySelector("#chat-thread-form img")'), false); + assert.equal(await b.evaluate('window.injected'), undefined); + await set('#chat-thread-name', 'Bean discussion'); + await click('#chat-thread-form button'); + assert.equal(await b.evaluate('window.__writes.length'), 0); + await set('#chat-instance', 'bean'); + await refresh(); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), 'bean'); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-name").value'), 'Bean discussion'); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), false); + await b.evaluate(`window.__answers.push({body:{id:'${A}',instance:'bean'}});true`); + await click('#chat-thread-form button'); + await wait('window.__writes.length===1'); + assert.deepEqual(await b.evaluate('window.__writes[0]'), { path: '/api/chat/threads', body: { project: P, name: 'Bean discussion', instance: 'bean' }, redirect: 'manual' }); +}); + +test('C2-09a picker: metadata failure, empty set and removed pick explain disabled state and recover without switching', async t => { + const { b, wait, set, refresh } = await setup(t); + await b.evaluate(`window.__instanceFailure=true;location.hash='#/chat/projects/${P}';true`); + await wait('document.querySelector("#chat-thread-form") && !document.querySelector(".chat-loading")'); + assert.match(await b.evaluate('document.querySelector("#chat-thread-form [role=status]").textContent'), /read the thread instances/); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); + await b.evaluate('window.__instanceFailure=false;window.__instances=[];true'); + await refresh(); + assert.match(await b.evaluate('document.querySelector("#chat-thread-form [role=status]").textContent'), /No thread instances/); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); + await b.evaluate("window.__instances=[{instance:'bean',name:'Bean'},{instance:'chet',name:'Chet'}];true"); + await refresh(); + await set('#chat-instance', 'bean'); + await b.evaluate("window.__instances=[{instance:'chet',name:'Chet'}];true"); + await refresh(); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), ''); + assert.match(await b.evaluate('document.querySelector("#chat-thread-form [role=status]").textContent'), /no longer available/); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); + await refresh(); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), '', 'another refresh must not silently select the one remaining instance'); + await set('#chat-instance', 'chet'); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), false); +}); + +test('C2-09a picker: a single instance is chosen explicitly but read-only mode cannot create a thread', async t => { + const { b, wait } = await setup(t, '?readonly'); + await b.evaluate(`location.hash='#/chat/projects/${P}';true`); + await wait('document.querySelectorAll("#chat-instance option").length===2'); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").value'), 'assistant'); + assert.equal(await b.evaluate('document.querySelector("#chat-instance").disabled'), true); + assert.equal(await b.evaluate('document.querySelector("#chat-thread-form button").disabled'), true); + assert.match(await b.evaluate('document.querySelector(".chat-area").textContent'), /only reads/); + assert.equal(await b.evaluate('window.__writes.length'), 0); +}); diff --git a/packages/webui/tests/chat-instances.test.mjs b/packages/webui/tests/chat-instances.test.mjs new file mode 100644 index 000000000..893c38f80 --- /dev/null +++ b/packages/webui/tests/chat-instances.test.mjs @@ -0,0 +1,87 @@ +import test from 'node:test'; +import assert from 'node:assert/strict'; +import { request } from 'node:http'; +import { mkdtempSync, rmSync } from 'node:fs'; +import { join } from 'node:path'; +import { tmpdir } from 'node:os'; +import { startServer } from '../src/serve.mjs'; +import { startTesterSocket, TESTER_ORIGIN } from '../src/tester-socket.mjs'; +import { close } from './fixture.mjs'; + +const path = '/api/chat/instances'; +const preview = { host: 'chat.example', header: 'x-user', user: 'jason' }; +const snapshot = { + revision: 'a'.repeat(64), plugins: { private: 'do-not-project' }, + instances: { + assistant: { definition: 'assistant', vars: { model: 'private-model' } }, + bean: { definition: 'companion', vars: { 'persona.name': 'Bean', 'persona.sha256': 'private-hash' } }, + chet: { definition: 'continuity', vars: { 'persona.name': 'Bean', file: '/private/path' } }, + pm: { definition: 'pm', vars: { 'persona.name': 'Hidden PM' } }, + }, +}; +function get(server, { route = path, method = 'GET', headers = {} } = {}) { + const address = server.address(), unix = typeof address === 'string'; + return new Promise((resolve, reject) => { + const req = request({ ...(unix ? { socketPath: address } : { host: '127.0.0.1', port: address.port }), + path: route, method, headers: { host: unix ? new URL(TESTER_ORIGIN).host : preview.host, ...headers } }, res => { + let text = ''; + res.on('data', chunk => text += chunk); + res.on('end', () => resolve({ status: res.statusCode, text, value: JSON.parse(text) })); + }); + req.on('error', reject); req.end(); + }); +} +test('C2-09a metadata uses the human listener bus, projects labels only and refuses unauthenticated reads', async t => { + let calls = 0; + const server = await startServer({ port: 0, preview, bus: { instances: async () => { calls++; return snapshot; } } }); + t.after(() => close(server)); + const refused = await get(server); + assert.equal(refused.status, 401); + assert.equal(calls, 0); + const answer = await get(server, { headers: { 'x-user': 'jason' } }); + assert.equal(answer.status, 200); + assert.deepEqual(answer.value.rows, [{ instance: 'assistant', name: 'assistant' }, { instance: 'bean', name: 'Bean' }, { instance: 'chet', name: 'Bean' }]); + for (const row of answer.value.rows) assert.deepEqual(Object.keys(row).sort(), ['instance', 'name']); + assert.doesNotMatch(answer.text, /private|Hidden PM/); + assert.equal(calls, 1); + for (const options of [{ route: path + '?extra=1' }, { method: 'POST' }]) { + const result = await get(server, { ...options, headers: { 'x-user': 'jason' } }); + assert.ok([400, 405].includes(result.status)); + } + assert.equal(calls, 1, 'invalid metadata requests never call a bus'); +}); +test('C2-09a metadata uses the private tester listener own bus and no human transport', async t => { + const root = mkdtempSync(join(tmpdir(), 'chat219-')); t.after(() => rmSync(root, { recursive: true, force: true })); + let reads = 0; + const server = await startTesterSocket({ dataRoot: root, board: 'http://127.0.0.1:9', + bus: { instances: async () => { reads++; return { instances: { quill: { definition: 'proofreader', vars: { 'persona.name': 'Quill' } } } }; } } }); + t.after(() => close(server)); + const answer = await get(server); + assert.equal(answer.status, 200); + assert.deepEqual(answer.value.rows, [{ instance: 'quill', name: 'Quill' }]); + assert.equal(reads, 1); +}); +test('C2-09a metadata refuses missing or malformed reads and never echoes untrusted fields', async t => { + for (const bus of [null, { instances: async () => ({ instances: [] }) }, { instances: async () => { throw new Error('private-path'); } }]) { + const server = await startServer({ port: 0, preview, bus }); t.after(() => close(server)); + const answer = await get(server, { headers: { 'x-user': 'jason' } }); + assert.ok(answer.status >= 400); + assert.doesNotMatch(answer.text, /private-path/); + } +}); + +test('C2-09a metadata stays behind the SSO session check and a GET never refreshes session activity', async t => { + let session = null, reads = 0; + const touches = []; + const sso = { publicHost: 'sso-chat.example', publicOrigin: 'https://sso-chat.example', formTargets: () => [], + session: (_req, { touch }) => { touches.push(touch); return session; } }; + const server = await startServer({ port: 0, sso, bus: { instances: async () => { reads++; return snapshot; } } }); + t.after(() => close(server)); + const headers = { host: sso.publicHost }; + assert.equal((await get(server, { headers })).status, 401); + assert.equal(reads, 0); + session = { name: 'jason' }; + assert.equal((await get(server, { headers })).status, 200); + assert.equal(reads, 1); + assert.deepEqual(touches, [false, false]); +}); diff --git a/packages/webui/tests/chat.test.mjs b/packages/webui/tests/chat.test.mjs index 8c11611cf..b47ecf1a7 100644 --- a/packages/webui/tests/chat.test.mjs +++ b/packages/webui/tests/chat.test.mjs @@ -335,3 +335,21 @@ test('the chat seam names the transport it is given: the human one and the teste assert.throws(() => createChat({}), { name: 'TypeError', message: 'chat needs a bus transport function' }); assert.equal(typeof createChat({ call: async () => ({}) }).projects, 'function'); }); + +test('C2-09a createThread forwards an explicit instance through the existing channel', async t => { + const c = await setup(t); + c.state.answer = { id: T, instance: 'bean' }; + const answer = await c.get('/api/chat/threads', { method: 'POST', value: { project: P, name: 'Bean work', instance: 'bean' } }); + assert.equal(answer.status, 200); + assert.deepEqual(c.calls, [['thread.create', { project: P, name: 'Bean work', instance: 'bean' }]]); + assert.equal(answer.data.instance, 'bean'); +}); +test('C2-09a malformed instance and extra thread fields refuse before the channel write', async t => { + const c = await setup(t); + for (const value of [null, 7, {}, [], '', 'bad name', 'bean/../../other']) { + const answer = await c.get('/api/chat/threads', { method: 'POST', value: { project: P, name: 'work', instance: value } }); + assert.equal(answer.status, 400, JSON.stringify(value)); + } + assert.equal((await c.get('/api/chat/threads', { method: 'POST', value: { project: P, name: 'work', instance: 'bean', extra: true } })).status, 400); + assert.deepEqual(c.calls, []); +});