mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-20 00:53:46 +01:00
## Thinking Path > - Paperclip manages work for companies of AI agents. > - App writes can require a human review for each invocation. > - Decline and Always allow are separate controls. > - The gateway currently reuses a declined action across later human requests with the same arguments. > - A new request can therefore receive the old decline instead of a new review. > - This change uses the saved run and human comment to distinguish a new intent. > - The new intent still follows the current policy and approval gates. ## Linked Issues or Issue Description **What happened?** After a person declines an Ask first app write, a later request from that person for the same write receives `action_declined`. The gateway matches company, task, agent, tool and arguments across runs. It does not distinguish the later human request. **Expected behavior** A distinct later request from the same responsible person can open a new Ask first review. It must not execute the write or reuse approval. A retry of the declined intent, including its rejection continuation, must remain declined. **Steps to reproduce** 1. Ask an agent to perform an app write that requires approval. 2. Decline the proposed action. 3. Send a new request for the same action. 4. Observe that the gateway reuses the old rejection and creates no new pending review. **Paperclip version or commit** The behavior is present on master `7b8aa60b6d` and in a retained native journey. The focused parent control reproduces the matching error. No related open fix was found in the scoped PR search. ## What Changed - Only a rejected match can be released for a later human intent. - Require the active run, company, task, agent and responsible person to match the saved direct-human comment. Reject derived, deleted, untrusted or system-authored input and invalid chronology. - Preserve all other replay states, signed argument checks, current policy, unresolved-effect guards, approval and execution owners. - Document the distinction between decline and a standing policy change. ## Verification - Actual matching-function control: parent fails; this change passes. The control uses the unchanged owner body with synthetic database rows. It does not claim real PostgreSQL or provider verification. - 52 focused pure controls pass. These cover same-intent retries, rejection continuations, foreign scope and actor, invalid authorship, chronology and other action states. - TypeScript 7.0.2 checks pass for the new helper and its owning tests. - `node scripts/check-maps.mjs` and `git diff --check` pass. - The existing gateway service suite has a new signed-action integration test. It covers ordinary and reopened human requests. It requires the old rejection plus one new pending request, a stable retry, and no completed call. Each leaf prints one fixed success label after its assertions so CI can distinguish execution from a platform skip. No local PostgreSQL, app, model or provider execution was run. Current-head normal CI41594912212 completed with48 successful jobs and1 expected skip. Both real PostgreSQL leaves executed and emitted their post-assertion markers; the owning job passed. Full typecheck/build checks passed in CI. ## Risks This adds one indexed run/comment read only after a rejected matching action. Missing identity or ambiguous provenance keeps the old rejection. Legacy and external comments without direct-human provenance do not reset it. A new human intent is not consent to execute. This core PR remains unmerged. Candidate native journey qualification is separate; the original failed journey is retained. ## Model Used OpenAI `gpt-6.1-sol` with reasoning, source inspection and tool use. The host did not expose the context-window size. ## 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 #123` / `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 - [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>
83 lines
6.2 KiB
TypeScript
83 lines
6.2 KiB
TypeScript
import { afterEach, expect, it, vi } from "vitest";
|
|
import { isNewHumanIntentAfterToolRejection } from "./tool-action-replay-intent.js";
|
|
afterEach(() => vi.useRealTimers());
|
|
function fixture() {
|
|
vi.useFakeTimers(); vi.setSystemTime(10_000);
|
|
return {
|
|
session: { companyId: "company", issueId: "task", agentId: "agent", runId: "new-run", responsibleUserId: "person" },
|
|
rejection: { companyId: "company", issueId: "task", status: "rejected", requestedByAgentId: "agent",
|
|
resolvedByUserId: "person", decidedByUserId: "person", resolvedByAgentId: null, decidedByAgentId: null,
|
|
resolvedAt: new Date(2_000), decidedAt: new Date(2_000) },
|
|
invocation: { companyId: "company", issueId: "task", agentId: "agent", runId: "original-run" },
|
|
run: { id: "new-run", companyId: "company", agentId: "agent", status: "running", responsibleUserId: "person", startedAt: new Date(4_000),
|
|
issueId: "task", nativeIssueId: "task", contextSnapshot: { source: "issue.comment", issueId: "task", taskId: "task", commentId: "new-human" } },
|
|
comment: { id: "new-human", companyId: "company", issueId: "task", authorType: "user", authorUserId: "person", authorAgentId: null, createdByRunId: null,
|
|
derivedAuthorAgentId: null, derivedCreatedByRunId: null, derivedAuthorSource: null, deletedAt: null, sourceTrust: null, presentation: null,
|
|
createdAt: new Date(3_000), body: "Please try that update again and let me review it." },
|
|
} satisfies Parameters<typeof isNewHumanIntentAfterToolRejection>[0];
|
|
}
|
|
it("distinguishes a new authenticated human request after rejection without approving it", () => {
|
|
const f = fixture(); expect(isNewHumanIntentAfterToolRejection(f)).toBe(true);
|
|
expect(f.rejection.status).toBe("rejected");
|
|
});
|
|
it("accepts the ordinary direct-human reopen source with the same persisted intent checks", () => {
|
|
const f = fixture(); f.run.contextSnapshot.source = "issue.comment.reopen";
|
|
expect(isNewHumanIntentAfterToolRejection(f)).toBe(true);
|
|
});
|
|
it.each(["automatic", "derived", "forged comment"])("does not treat %s reopen input as a new human intent", fault => {
|
|
const f = fixture(); f.run.contextSnapshot.source = "issue.comment.reopen";
|
|
if (fault === "automatic") f.run.contextSnapshot.source = "automation.issue.comment.reopen";
|
|
else if (fault === "derived") Object.assign(f.comment, { derivedCreatedByRunId: "agent-run" });
|
|
else f.run.contextSnapshot.commentId = "unrelated-human-comment";
|
|
expect(isNewHumanIntentAfterToolRejection(f)).toBe(false);
|
|
});
|
|
const invalid: Array<[string, (f: ReturnType<typeof fixture>) => void]> = [
|
|
["same intent retry", f => { f.comment.createdAt = new Date(1_000); }],
|
|
["same run retry", f => { f.invocation.runId = f.run.id; }],
|
|
["rejection continuation", f => { f.run.contextSnapshot.source = "tool_action_review"; }],
|
|
["automatic continuation", f => { f.run.contextSnapshot.source = "chat_task_completed"; }],
|
|
["foreign session company", f => { f.session.companyId = "foreign"; }],
|
|
["foreign request company", f => { f.rejection.companyId = "foreign"; }],
|
|
["foreign request task", f => { f.rejection.issueId = "foreign"; }],
|
|
["foreign invocation actor", f => { f.invocation.agentId = "foreign"; }],
|
|
["foreign invocation task", f => { f.invocation.issueId = "foreign"; }],
|
|
["foreign invocation company", f => { f.invocation.companyId = "foreign"; }],
|
|
["foreign run company", f => { f.run.companyId = "foreign"; }],
|
|
["foreign run actor", f => { f.run.agentId = "foreign"; }],
|
|
["stopped run", f => { f.run.status = "cancelled"; }],
|
|
["foreign context source", f => { f.run.contextSnapshot.issueId = "foreign"; }],
|
|
["contradictory task binding", f => { f.run.contextSnapshot.taskId = "foreign"; }],
|
|
["foreign native task", f => { f.run.nativeIssueId = "foreign"; }],
|
|
["foreign stored run task", f => { f.run.issueId = "foreign"; }],
|
|
["wrong canonical comment", f => { f.run.contextSnapshot.commentId = "old-human"; }],
|
|
["foreign comment task", f => { f.comment.issueId = "foreign"; }],
|
|
["foreign comment company", f => { f.comment.companyId = "foreign"; }],
|
|
["another human", f => { f.comment.authorUserId = "another-person"; }],
|
|
["wrong responsible user", f => { f.run.responsibleUserId = "another-person"; }],
|
|
["another person's rejection", f => { f.rejection.decidedByUserId = "another-person"; }],
|
|
["unknown rejection author", f => { f.rejection.decidedByUserId = ""; f.rejection.resolvedByUserId = ""; }],
|
|
["agent-attributed comment", f => { Object.assign(f.comment, { authorAgentId: "agent" }); }],
|
|
["agent-created comment", f => { Object.assign(f.comment, { createdByRunId: "old-run" }); }],
|
|
["derived agent comment", f => { Object.assign(f.comment, { derivedAuthorAgentId: "agent" }); }],
|
|
["derived run comment", f => { Object.assign(f.comment, { derivedCreatedByRunId: "run" }); }],
|
|
["derived attribution", f => { Object.assign(f.comment, { derivedAuthorSource: "run_log" }); }],
|
|
["deleted human comment", f => { Object.assign(f.comment, { deletedAt: new Date(5_000) }); }],
|
|
["untrusted external source", f => { Object.assign(f.comment, { sourceTrust: { disposition: "quarantined" } }); }],
|
|
["system notice", f => { Object.assign(f.comment, { presentation: { kind: "system_notice" } }); }],
|
|
["non-user author", f => { f.comment.authorType = "agent"; }],
|
|
["empty comment", f => { f.comment.body = " "; }],
|
|
["new-session command", f => { f.comment.body = "/new"; }],
|
|
["equal rejection time", f => { f.comment.createdAt = new Date(2_000); }],
|
|
["comment after run started", f => { f.comment.createdAt = new Date(5_000); }],
|
|
["future run", f => { f.run.startedAt = new Date(11_000); }],
|
|
["invalid decision time", f => { f.rejection.decidedAt = new Date(NaN); }],
|
|
["invalid comment time", f => { f.comment.createdAt = new Date(NaN); }],
|
|
["invalid start time", f => { f.run.startedAt = new Date(NaN); }],
|
|
];
|
|
it.each(invalid)("preserves the prior rejection for %s", (_name, mutate) => {
|
|
const f = fixture(); mutate(f); expect(isNewHumanIntentAfterToolRejection(f)).toBe(false);
|
|
});
|
|
it.each(["pending", "approved", "executing", "executed", "failed", "cancelled"])("never bypasses %s action state", status => {
|
|
const f = fixture(); f.rejection.status = status; expect(isNewHumanIntentAfterToolRejection(f)).toBe(false);
|
|
});
|