mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-16 11:59:26 +01:00
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The task composer holds messages that an agent has not consumed yet. > - A board operator needs to remove an agent-authored pending message before it starts more work. > - The queue currently grants discard only to the human author, so these messages have a disabled trash button and the API rejects the board user. > - This change derives the board discard permission in the queue snapshot and uses that same permission under the existing issue lock. > - The route checks task access before it changes the queue. The benefit is a working discard action with the existing company, privacy, revision, and dispatch guards. ## Linked Issues or Issue Description Fixes #15464. Supersedes #15465 with an independent implementation and red/green regression evidence. Thanks to @MMTE for identifying the gap and proposing a fix. Related: #18958 addresses deletion of comments that have already been delivered; this change covers pending queue entries. ## What Changed - Queue responses grant board users discard permission for agent-authored messages. Editing and other human authors' messages retain their existing permissions. - The discard use case honors the actor-specific permission rebuilt under the queue lock. Author agents keep their existing cancellation path. - The board discard route checks private-task access before deletion. The denial for a non-author agent now points to the board action instead of saying that no board action exists. - Tests cover queue states, cancellation of an empty continuation, persistence, queue linkage, audit attribution, private-task denial, and cross-company denial. The API reference describes the supported action. ## Verification - Red before the fix: the two snapshot regressions returned `canDiscard: false`; the discard regression threw `QueuedCommentMutationForbiddenError`. The route regressions also caught the disabled permission, missing private-task guard, and stale agent denial text. - Green after the fix: 30 queue snapshot and use-case tests; all five focused route regressions with real disposable Postgres; six Postgres adapter tests; 42 existing queue UI tests. - Server typecheck: `cd server && node scripts/typecheck-layers.mjs` passes. `node scripts/check-module-boundaries.mjs` and `git diff --check` pass. - The [complete PR workflow](https://github.com/paperclipai/paperclip/actions/runs/43484735280) passes on head `164200602828e23f01a8b209ff954821c5e14e9a`. [Greptile reports 5/5](https://github.com/paperclipai/paperclip/pull/21458#issuecomment-6861824451) on that same head with no open review threads. - Local scope: the 83 focused snapshot, use-case, route, adapter, and UI checks pass. The existing queue UI suite has 42 tests. The new route cases use real disposable Postgres and verify the resulting database state after reload. - Additional local full-suite attempt: 174 tests passed and 13 approval-related tests failed with cleanup timeouts, Postgres deadlocks, and duplicate-prefix errors after cleanup failed. The first failure, `queues request_confirmation resolution while its legacy source run is still running`, also fails on unchanged base code with the same 30-second `afterEach` and `afterAll` timeouts. This broader local run is not counted as a pass. No test timeout or assertion was weakened. CI supplies the broad passing result for the committed head. - `node scripts/local-verify.mjs --explain` returned exit 2 because the API documentation is a global CI input; all of its required CI lanes passed. The initial adapter attempt could not initialize local Postgres under the sandbox; the same six checks passed with local-process access. The initial cross-company fixture was corrected to use a distinct company prefix. ## Risks - This grants board operators a new action on agent-authored pending messages. The route must pass company membership and private-task access checks; the mutation retains its transaction, revision, and dispatch protections. - No schema or migration is required. The UI uses its existing discard control and error handling. - The staging task was inspected to confirm the cause. No staging data was deleted, and a deployed UI retest is not claimed by these local and CI checks. ## Model Used - OpenAI Codex, GPT-6, with reasoning, repository inspection, tool use, and code execution. The exact runtime model ID and context limit are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass — the focused regression, adapter, and UI checks listed above; the additional broad local cleanup failure is documented separately - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge Co-authored-by: Paperclip <noreply@paperclip.ing>
179 lines
7.4 KiB
TypeScript
179 lines
7.4 KiB
TypeScript
import { describe, expect, it } from "vitest";
|
|
import { buildQueuedCommentQueueSnapshot, decideQueuedCommentQueueSteering, withQueuedCommentIdsInRunContext } from "./issue-queued-comment-queue.js";
|
|
|
|
describe("decideQueuedCommentQueueSteering", () => {
|
|
it("answers unsupported on the legacy protocol", () => {
|
|
const decision = decideQueuedCommentQueueSteering({
|
|
state: "deferred",
|
|
queueRunRuntimeMode: null,
|
|
activeRun: { id: "run-1", runtimeMode: "legacy" },
|
|
assignedAgentAdapterType: "codex_local",
|
|
queuedCommentCount: 1,
|
|
});
|
|
|
|
expect(decision).toEqual({ protocol: "legacy", kind: "unsupported" });
|
|
});
|
|
|
|
it("answers temporarily_unavailable for a promoted native queue with no deferred run", () => {
|
|
const decision = decideQueuedCommentQueueSteering({
|
|
state: "queued",
|
|
queueRunRuntimeMode: "native",
|
|
activeRun: null,
|
|
assignedAgentAdapterType: "paperclip_runner",
|
|
queuedCommentCount: 1,
|
|
});
|
|
|
|
expect(decision).toEqual({ protocol: "paperclip_runner_v1", kind: "temporarily_unavailable" });
|
|
});
|
|
|
|
it("answers temporarily_unavailable when the queue holds no live comments", () => {
|
|
const decision = decideQueuedCommentQueueSteering({
|
|
state: "deferred",
|
|
queueRunRuntimeMode: null,
|
|
activeRun: { id: "run-1", runtimeMode: "native" },
|
|
assignedAgentAdapterType: "paperclip_runner",
|
|
queuedCommentCount: 0,
|
|
});
|
|
|
|
expect(decision).toEqual({ protocol: "paperclip_runner_v1", kind: "temporarily_unavailable" });
|
|
});
|
|
|
|
it("tells the caller it may probe a running deferred turn on the native protocol", () => {
|
|
const decision = decideQueuedCommentQueueSteering({
|
|
state: "deferred",
|
|
queueRunRuntimeMode: null,
|
|
activeRun: { id: "run-1", runtimeMode: "native" },
|
|
assignedAgentAdapterType: "paperclip_runner",
|
|
queuedCommentCount: 1,
|
|
});
|
|
|
|
expect(decision).toEqual({ protocol: "paperclip_runner_v1", kind: "probe", steeringRunId: "run-1" });
|
|
});
|
|
|
|
// Acceptance-criterion fact pattern: a deferred queue whose active run
|
|
// has not resolved a runtime mode yet, for an agent on the
|
|
// `paperclip_runner` adapter. The protocol resolves to
|
|
// `paperclip_runner_v1` through the adapter-type fallback, and the
|
|
// decision hands the run to the caller to probe live — it never answers
|
|
// the flat "unsupported" value a duplicated, unshared rule can drift to.
|
|
it("resolves the protocol through the adapter-type fallback and asks the caller to probe", () => {
|
|
const decision = decideQueuedCommentQueueSteering({
|
|
state: "deferred",
|
|
queueRunRuntimeMode: null,
|
|
activeRun: { id: "run-1", runtimeMode: null },
|
|
assignedAgentAdapterType: "paperclip_runner",
|
|
queuedCommentCount: 1,
|
|
});
|
|
|
|
expect(decision).toEqual({ protocol: "paperclip_runner_v1", kind: "probe", steeringRunId: "run-1" });
|
|
});
|
|
|
|
it("keeps a preparing Paperclip Runner on the steering protocol until runtime selection finishes", () => {
|
|
expect(decideQueuedCommentQueueSteering({ state: "deferred", queueRunRuntimeMode: null,
|
|
activeRun: { id: "preparing-run", runtimeMode: "legacy", runtimeModeResolvedAt: null,
|
|
runnerProfileJson: { adapterDispatch: { adapterType: "paperclip_runner" } } },
|
|
assignedAgentAdapterType: "paperclip_runner", queuedCommentCount: 1,
|
|
})).toEqual({ protocol: "paperclip_runner_v1", kind: "temporarily_unavailable" });
|
|
});
|
|
|
|
it("preserves the historical adapter when mutable agent settings change", () => {
|
|
expect(decideQueuedCommentQueueSteering({ state: "deferred", queueRunRuntimeMode: null,
|
|
activeRun: { id: "legacy-run", runtimeMode: "legacy", runtimeModeResolvedAt: null,
|
|
runnerProfileJson: { adapterDispatch: { adapterType: "codex_local" } } },
|
|
assignedAgentAdapterType: "paperclip_runner", queuedCommentCount: 1,
|
|
})).toEqual({ protocol: "legacy", kind: "unsupported" });
|
|
});
|
|
});
|
|
|
|
describe("buildQueuedCommentQueueSnapshot entry permissions", () => {
|
|
const baseFacts = {
|
|
issueId: "issue-1",
|
|
queueId: "queue-1",
|
|
state: "queued" as const,
|
|
activeRunId: null,
|
|
protocol: "legacy" as const,
|
|
steeringDisposition: "unsupported" as const,
|
|
};
|
|
|
|
it("grants edit and discard to the user who authored the queued comment", () => {
|
|
const queue = buildQueuedCommentQueueSnapshot({
|
|
...baseFacts,
|
|
actorType: "user",
|
|
actorId: "user-1",
|
|
comments: [{ id: "comment-1", updatedAt: new Date(), authorUserId: "user-1", authorAgentId: null }],
|
|
});
|
|
|
|
expect(queue.entries[0]?.canEdit).toBe(true);
|
|
expect(queue.entries[0]?.canDiscard).toBe(true);
|
|
});
|
|
|
|
it("denies edit and discard to a user who did not author the queued comment", () => {
|
|
const queue = buildQueuedCommentQueueSnapshot({
|
|
...baseFacts,
|
|
actorType: "user",
|
|
actorId: "user-1",
|
|
comments: [{ id: "comment-1", updatedAt: new Date(), authorUserId: "user-2", authorAgentId: null }],
|
|
});
|
|
|
|
expect(queue.entries[0]?.canEdit).toBe(false);
|
|
expect(queue.entries[0]?.canDiscard).toBe(false);
|
|
});
|
|
|
|
it("denies edit and discard to an agent actor even when the comment carries a matching author id", () => {
|
|
const queue = buildQueuedCommentQueueSnapshot({
|
|
...baseFacts,
|
|
actorType: "agent",
|
|
actorId: "user-1",
|
|
comments: [{ id: "comment-1", updatedAt: new Date(), authorUserId: "user-1", authorAgentId: null }],
|
|
});
|
|
|
|
expect(queue.entries[0]?.canEdit).toBe(false);
|
|
expect(queue.entries[0]?.canDiscard).toBe(false);
|
|
});
|
|
|
|
it.each(["deferred", "queued"] as const)("lets the board discard an agent message in a %s queue without editing it", (state) => {
|
|
const queue = buildQueuedCommentQueueSnapshot({
|
|
...baseFacts,
|
|
state,
|
|
actorType: "user",
|
|
actorId: "user-1",
|
|
comments: [{ id: "comment-1", updatedAt: new Date(), authorUserId: null, authorAgentId: "agent-1" }],
|
|
});
|
|
|
|
expect(queue.entries[0]).toMatchObject({ canEdit: false, canDiscard: true });
|
|
});
|
|
});
|
|
|
|
describe("withQueuedCommentIdsInRunContext", () => {
|
|
it("invalidates every generated task projection and ownership metadata", () => {
|
|
const result = withQueuedCommentIdsInRunContext({
|
|
issueId: "issue-1",
|
|
preserved: "keep",
|
|
paperclipWake: { comments: [{ id: "comment-1" }] },
|
|
paperclipWakeComment: { id: "comment-1" },
|
|
paperclipTaskMarkdown: "historical",
|
|
paperclipTaskMarkdownCompact: "historical compact",
|
|
paperclipTaskMarkdownAssignment: "assignment",
|
|
paperclipTaskMarkdownAssignmentCompact: "assignment compact",
|
|
paperclipTurnContext: { version: 1, events: { owner: "wake_prompt" } },
|
|
}, ["comment-2"], ["source-wake"]);
|
|
|
|
expect(result).toMatchObject({
|
|
issueId: "issue-1",
|
|
preserved: "keep",
|
|
queuedCommentSourceWakeIds: ["source-wake"],
|
|
wakeCommentIds: ["comment-2"],
|
|
wakeCommentId: "comment-2",
|
|
commentId: "comment-2",
|
|
});
|
|
expect(result).not.toHaveProperty("paperclipWake");
|
|
expect(result).not.toHaveProperty("paperclipWakeComment");
|
|
expect(result).not.toHaveProperty("paperclipTaskMarkdown");
|
|
expect(result).not.toHaveProperty("paperclipTaskMarkdownCompact");
|
|
expect(result).not.toHaveProperty("paperclipTaskMarkdownAssignment");
|
|
expect(result).not.toHaveProperty("paperclipTaskMarkdownAssignmentCompact");
|
|
expect(result).not.toHaveProperty("paperclipTurnContext");
|
|
expect(withQueuedCommentIdsInRunContext(result, ["comment-3"]).queuedCommentSourceWakeIds).toEqual(["source-wake"]);
|
|
});
|
|
});
|