mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-14 00:26:28 +01:00
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - The chat-feature evals measure whether agents complete real chat requests through the product > - The chat catalog eval harness and its repairs live on the `chat-eval/master-candidate` branch, not on master > - Today's work made 139 of 242 chat evals pass in verified single-case live runs. Each product fix landed on master as its own pull request > - The eval-side changes (setups, graders, fault seams, evidence capture) still exist only on the candidate > - This pull request lands the harness on master. Master stays the source of truth for product behavior > - The benefit is one tree: evals run against master commits, and product fixes reach the evals as soon as they merge ## Linked Issues or Issue Description Refs #19441, Refs #19599, Refs #20270, Refs #20334, Refs #20342, Refs #20344, Refs #20354, Refs #20360, Refs #20373, Refs #20382, Refs #20409 **What existing behavior does this improve?** The chat catalog eval harness and the product it evaluates live on a long-running branch outside master. Eval fixes do not reach master, so master-based eval runs use old setups and graders. **Subsystem affected** `tests/runner-e2e` chat catalog harness, plus a small set of product changes that exist only on the candidate. **Proposed change** Land the harness on master. Every product fix from this eval work already merged to master as its own pull request. Conflicts keep master's product behavior. ## What Changed - Most of the diff (808 of 851 files) is under `tests/runner-e2e/`: chat-catalog domains, scenarios, descriptors, the TestOps manifest, private evidence tooling, and their tests. - Graders now check saved state, setup literals and rendered or clicked links. Many keyword and regex checks on free-form replies are removed. Each changed grader has a bumped `oracleVersion` and calibration tests (a valid reply, a reworded valid reply, and broken variants that must fail). - The shared link helper `hasLink` no longer treats a list link (for example `/COM/issues?view=mine`) as a link to every task without a URL key. - Setup fixes build real state through supported product routes: routine history, agent grants through the board grant route, budget holds with a real cap, and a storage-layer archive fault for `inbox-search.archive-partial-failure`. - Browser boundary journeys (sign-in, chat-disabled, onboarding) now write the standard pass evidence files. - Retargeted cases, approved by the maintainer: `agent-chat.chat-disabled` (Conference Room gate), `agents.restore-instructions` (no saved instruction history), `access.manage-membership` (member permission edits stay board-only). - The private evidence trusted files (`manual-private-handoff.ts`, `manual-private-retention.ts`, `manual-private-report-assets.ts`) are byte-identical to master. - Product files outside `tests/` (43 files, +1,328 −49) carry candidate-only changes that master does not yet have. Most are tests. The main code changes are in `request-human-input.ts`, `heartbeat-run-summary.ts`, `native-completion-feedback.ts`, `native-run-finalizer.ts`, `openapi.ts`, `secrets.ts` and `IssueChatLongThreadPerf.tsx`. ## Verification - `pnpm vitest run chat-catalog-domains chat-catalog-contract.test.ts chat-catalog-consistency.test.ts unit-launcher.test.mjs --testTimeout=180000` in `tests/runner-e2e`: 269 files pass, 3,550 tests pass. - `node --test chat-full-catalog-testops-manifest.test.mjs`: 10 pass. - `node tests/runner-e2e/source-sdk-public-inputs.mjs <HEAD> --check`: no change. - `tsc` for `server` and `tests/runner-e2e`: clean. - Live single-case verification: 139 of 242 cases passed on paid runs, and a person read the saved evidence for each. These are single-case runs, not one full-cohort run. ## Risks - Large diff. Most of it is test-only, but the 43 non-test files change product code paths. Review those files first. - Some graders are now less strict on reply wording. A person must read replies for meaning on those cases. - 42 cases need disposable outside accounts (Notion, Slack, GitHub, Daytona and others) and stay blocked. `access.invite-teammate` stays failing until local agent isolation lands, because the runner's durable state keeps a live invite token. - The paid workflow compares the trusted private-evidence files with master. They are unchanged, so this risk is low. ## Model Used Anthropic Claude Opus 5.5 (`claude-opus-5-5`), with tool use and code execution, in Claude Code. Subagents used the same model. ## 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 - [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 - [ ] All Paperclip CI gates are green - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Paperclip <noreply@paperclip.ing>
72 lines
8.8 KiB
TypeScript
72 lines
8.8 KiB
TypeScript
import type { ChatCatalogContext, ChatCatalogQueryBinding } from "./chat-catalog-contract.js";
|
|
|
|
const allowedCatalogTokens: Readonly<Record<string, readonly string[]>> = {
|
|
"access.other-company": ["BIRCH-42"], "projects.add-repository": ["acorn/site"],
|
|
"projects.replace-repository": ["acorn/site-old", "acorn/site-new"],
|
|
"workspaces.prepare-repository": ["acorn/site"], "workspaces.private-repository-no-access": ["acorn/site"],
|
|
"chat-channels.outbound-email":["reviewer@example.test"],
|
|
};
|
|
const allowedResourceKinds: Readonly<Record<string, string>> = { "access.other-company": "issue",
|
|
"projects.add-repository": "project-repository", "projects.replace-repository": "project-repository",
|
|
"workspaces.prepare-repository": "project-repository", "workspaces.private-repository-no-access": "project-repository","chat-channels.outbound-email":"agentmail-recipient" };
|
|
export interface RenderedChatQuery {
|
|
originalQuery: string; renderedQuery: string;
|
|
bindings: Array<{ catalogToken: string; fixtureValue: string; resourceKind: "issue" | "project-repository" | "agentmail-recipient"; resourceId: string | number; companyId: string; connectionId?: string }>;
|
|
}
|
|
const uuid=/^[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}$/i;
|
|
export function safeFixtureEmail(value:unknown):value is string{return typeof value==="string"&&value.length<=254&&!/[\r\n\s<>]/.test(value)&&!value.includes("..")&&/^[A-Za-z0-9.!#$%&'*+/=?^_`{|}~-]+@[A-Za-z0-9](?:[A-Za-z0-9.-]*[A-Za-z0-9])?\.[A-Za-z]{2,}$/.test(value);}
|
|
export function renderChatCatalogQuery(scenarioId: string, originalQuery: string, specs: readonly ChatCatalogQueryBinding[] = [], state: Record<string, unknown>): RenderedChatQuery {
|
|
let renderedQuery = originalQuery; const bindings: RenderedChatQuery["bindings"] = [];
|
|
for (const spec of specs) {
|
|
if (!allowedCatalogTokens[scenarioId]?.includes(spec.catalogToken) || spec.resourceKind !== allowedResourceKinds[scenarioId] || spec.valueField !== (spec.resourceKind === "issue" ? "identifier" : spec.resourceKind==="agentmail-recipient"?"email":"fullName")
|
|
|| bindings.some(row => row.catalogToken === spec.catalogToken) || originalQuery.split(spec.catalogToken).length !== 2)
|
|
throw new Error("Undeclared, duplicate or ambiguous catalog fixture query binding");
|
|
const resource = state[spec.stateKey] as { id?: unknown; companyId?: unknown; identifier?: unknown; fullName?: unknown; connectionId?: unknown;email?:unknown } | undefined;
|
|
if (!resource || typeof resource.companyId !== "string" || !resource.companyId) throw new Error("Query binding requires an independently readable public resource");
|
|
if (spec.resourceKind === "issue") {
|
|
if (typeof resource.id !== "string" || !resource.id || typeof resource.identifier !== "string" || !/^[A-Z]+-\d+$/.test(resource.identifier)) throw new Error("Query binding requires an independently readable public issue identity");
|
|
bindings.push({ catalogToken: spec.catalogToken, fixtureValue: resource.identifier, resourceKind: "issue", resourceId: resource.id, companyId: resource.companyId });
|
|
} else if(spec.resourceKind==="agentmail-recipient"){
|
|
if(typeof resource.id!=="string"||!uuid.test(resource.id)||typeof resource.connectionId!=="string"||!uuid.test(resource.connectionId)||!safeFixtureEmail(resource.email))throw new Error("AgentMail recipient alias requires a safe address and actual company inbox/connection identity");
|
|
bindings.push({catalogToken:spec.catalogToken,fixtureValue:resource.email,resourceKind:"agentmail-recipient",resourceId:resource.id,companyId:resource.companyId,connectionId:resource.connectionId});
|
|
} else {
|
|
if (!Number.isSafeInteger(resource.id) || Number(resource.id) <= 0 || typeof resource.connectionId !== "string" || !/^[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}$/i.test(resource.connectionId)
|
|
|| typeof resource.fullName !== "string" || !/^[A-Za-z0-9][A-Za-z0-9_.-]*\/[A-Za-z0-9][A-Za-z0-9_.-]*$/.test(resource.fullName)) throw new Error("Repository query binding requires a numeric provider identity and current connection");
|
|
bindings.push({ catalogToken: spec.catalogToken, fixtureValue: resource.fullName, resourceKind: "project-repository", resourceId: resource.id as number, companyId: resource.companyId, connectionId: resource.connectionId });
|
|
}
|
|
}
|
|
// Apply positions from the original query, so a real identity cannot be rewritten as a later alias.
|
|
for (const binding of [...bindings].sort((a, b) => originalQuery.indexOf(b.catalogToken) - originalQuery.indexOf(a.catalogToken))) {
|
|
const index = originalQuery.indexOf(binding.catalogToken);
|
|
renderedQuery = renderedQuery.slice(0, index) + binding.fixtureValue + renderedQuery.slice(index + binding.catalogToken.length);
|
|
}
|
|
return { originalQuery, renderedQuery, bindings };
|
|
}
|
|
export async function resolveChatCatalogQuery(ctx: ChatCatalogContext, specs: readonly ChatCatalogQueryBinding[] | undefined, state: Record<string, unknown>) {
|
|
const rendered = renderChatCatalogQuery(ctx.scenario.id, ctx.scenario.query, specs, state);
|
|
for (const binding of rendered.bindings) {
|
|
if (binding.resourceKind === "issue") {
|
|
const saved = await ctx.input.api.get<{ id: string; companyId: string; identifier: string }>(`/api/issues/${binding.resourceId}`);
|
|
if (saved.id !== binding.resourceId || saved.companyId !== binding.companyId || saved.identifier !== binding.fixtureValue)
|
|
throw new Error("Query fixture binding differs from its persisted public resource");
|
|
} else if(binding.resourceKind==="agentmail-recipient"){
|
|
if(!ctx.input.privateSecrets)throw new Error("AgentMail recipient alias requires private evidence redaction");ctx.input.privateSecrets.push(binding.fixtureValue);
|
|
const config=ctx.input.privateInputs?.configuration("external-provider") as {providers?:{agentmail?:{disposable?:boolean;resourceIds?:string[];prerequisites?:Record<string,string>}}}|undefined,provider=config?.providers?.agentmail,p=provider?.prerequisites;
|
|
if(provider?.disposable!==true||p?.recipientAuthorization!=="self-addressed-only"||p.recipient!==binding.fixtureValue||p.inboxId!==binding.fixtureValue||!provider.resourceIds?.includes(p.inboxId)||binding.companyId!==ctx.input.fixtures.company.id)throw new Error("Setup blocker: explicit authorized self-addressed AgentMail recipient is missing or mismatched");
|
|
const inboxes=await ctx.input.api.get<Array<{id:string;companyId:string;connectionId:string;assignedAgentId:string;address:string|null;status:string}>>(`/api/companies/${binding.companyId}/email/inboxes`);
|
|
const endpoint=await ctx.input.api.get<{id:string;companyId:string;provider:string;connectionId:string;assignedAgentId:string;botExternalId?:string;status:string}>(`/api/chat-endpoints/${binding.resourceId}`);
|
|
const connection=await ctx.input.api.get<{id:string;companyId:string;status:string;enabled:boolean;config?:{provider?:string}}>(`/api/tool-connections/${binding.connectionId}`);
|
|
if(inboxes.filter(row=>row.id===binding.resourceId&&row.companyId===binding.companyId&&row.connectionId===binding.connectionId&&row.address===binding.fixtureValue&&row.status==="active"&&row.assignedAgentId===ctx.input.fixtures.agent.id).length!==1||endpoint.id!==binding.resourceId||endpoint.companyId!==binding.companyId||endpoint.provider!=="agentmail"||endpoint.connectionId!==binding.connectionId||endpoint.assignedAgentId!==ctx.input.fixtures.agent.id||endpoint.botExternalId!==binding.fixtureValue||endpoint.status!=="active"||connection.id!==binding.connectionId||connection.companyId!==binding.companyId||connection.status!=="active"||!connection.enabled||connection.config?.provider!=="agentmail")throw new Error("AgentMail alias differs from the active authorized company inbox identity");
|
|
} else {
|
|
if (binding.companyId !== ctx.input.fixtures.company.id) throw new Error("Repository binding belongs to another company");
|
|
const discovery = await ctx.input.api.get<{ repositories: Array<{ id: string; fullName: string; connectionIds?: string[] }> }>(`/api/companies/${binding.companyId}/project-repositories`);
|
|
const matches = discovery.repositories.filter(row => /^\d+$/.test(row.id) && Number(row.id) === binding.resourceId
|
|
&& row.fullName === binding.fixtureValue && row.connectionIds?.includes(binding.connectionId!));
|
|
const connection = await ctx.input.api.get<{ id: string; companyId: string; status: string; enabled: boolean; config?: { sourceTemplateKey?: string } }>(`/api/tool-connections/${binding.connectionId}`);
|
|
if (matches.length !== 1 || connection.id !== binding.connectionId || connection.companyId !== binding.companyId
|
|
|| connection.status !== "active" || !connection.enabled || connection.config?.sourceTemplateKey !== "github") throw new Error("Repository alias differs from active company-scoped provider discovery");
|
|
}
|
|
}
|
|
return rendered;
|
|
}
|