mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-13 15:55:46 +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>
109 lines
7.0 KiB
TypeScript
109 lines
7.0 KiB
TypeScript
import { CHAT_NATIVE_TIMEOUT_POLICY } from "./chat-native-timeout-policy.js";
|
|
import { createHash } from "node:crypto";
|
|
import { readFileSync, readdirSync } from "node:fs";
|
|
import { parseDocument } from "yaml";
|
|
import { chatCatalogCases } from "./chat-catalog-registry.js";
|
|
import { chatCatalogScenarios } from "./chat-catalog-inventory.js";
|
|
import type { RunnerProfileFixture, RunnerTaskFixture } from "./types.js";
|
|
|
|
export const CHAT_FEATURE_SUITE = "chat-features";
|
|
export const CHAT_FEATURE_BUDGET_CENTS = 500;
|
|
export const CHAT_FEATURE_GRADER = "paperclip.chat-features.v2";
|
|
export const CHAT_FEATURE_OPENCODE_MODEL = "openrouter/deepseek/deepseek-v4.1-flash-1017";
|
|
export const CHAT_FEATURE_CASES = ["create-backlog", "delegate-announcement"] as const;
|
|
export type ChatFeatureCase = typeof CHAT_FEATURE_CASES[number];
|
|
export interface ChatScenario {
|
|
id: string; capabilities: string[]; context: string; query: string; expected_outcome: string;
|
|
}
|
|
|
|
export function readChatFeatureScenarios(text = readFileSync(new URL("./chat-scenarios/tasks.yaml", import.meta.url), "utf8")): ChatScenario[] {
|
|
const document = parseDocument(text, { uniqueKeys: true });
|
|
if (document.errors.length) throw new Error(`Invalid chat scenario YAML: ${document.errors[0]!.message}`);
|
|
const value = document.toJS();
|
|
if (value?.schema_version !== 1 || value.feature !== "tasks" || value.status !== "planned" || !Array.isArray(value.scenarios)) {
|
|
throw new Error("Unsupported chat scenario catalog");
|
|
}
|
|
const ids = new Set<string>();
|
|
for (const scenario of value.scenarios) {
|
|
if (!scenario || ["id", "context", "query", "expected_outcome"].some(key => typeof scenario[key] !== "string" || !scenario[key].trim())
|
|
|| !Array.isArray(scenario.capabilities) || !scenario.capabilities.length
|
|
|| scenario.capabilities.some((id: unknown) => typeof id !== "string" || !/^[a-z0-9-]+\/[a-z0-9-]+$/.test(id))
|
|
|| ids.has(scenario.id)) throw new Error("Invalid or duplicate chat scenario");
|
|
ids.add(scenario.id);
|
|
}
|
|
return CHAT_FEATURE_CASES.map(caseId => {
|
|
const scenario = value.scenarios.find((candidate: ChatScenario) => candidate.id === `tasks.${caseId}`);
|
|
if (!scenario) throw new Error(`Missing chat scenario: tasks.${caseId}`);
|
|
return scenario;
|
|
});
|
|
}
|
|
export const chatFeatureScenarios = readChatFeatureScenarios();
|
|
function definitionSources(directory = new URL("./", import.meta.url)): string[] {
|
|
return readdirSync(directory, { withFileTypes: true }).sort((a, b) => a.name.localeCompare(b.name)).flatMap(entry => {
|
|
if (entry.isDirectory()) return ["node_modules", "results", "test-results", "playwright-report", ".results"].includes(entry.name)
|
|
? [] : definitionSources(new URL(`${entry.name}/`, directory));
|
|
return /\.(ts|mjs|yaml|json)$/.test(entry.name) && !entry.name.endsWith(".test.ts") && !entry.name.startsWith("chat-campaign")
|
|
? [new URL(entry.name, directory).href] : [];
|
|
});
|
|
}
|
|
// Includes every domain, helper, matcher and fixture source in this Product E2E family.
|
|
export const chatFeatureDefinitionDigest = createHash("sha256").update([
|
|
...definitionSources(), new URL("../../server/src/services/agent-conversations.ts", import.meta.url).href,
|
|
].map(file => {
|
|
const relative = new URL(file).pathname.replace(new URL("../../", import.meta.url).pathname, "");
|
|
const body = readFileSync(new URL(file), "utf8");
|
|
return JSON.stringify([relative, body]);
|
|
}).join("\n")).digest("hex");
|
|
|
|
export function chatFeatureProfile(profile: RunnerProfileFixture): RunnerProfileFixture {
|
|
return { ...profile, buildAgent(input) {
|
|
const { instructionsBundle: _manual, ...payload } = profile.buildAgent(input);
|
|
const config = payload.adapterConfig as Record<string, unknown>;
|
|
const { codexPermissionMode: _permission, ...adapterConfig } = config;
|
|
const throughputRouting = profile.id === "runner-opencode" && profile.provider === "opencode"
|
|
&& profile.model === CHAT_FEATURE_OPENCODE_MODEL && profile.adapterType === "paperclip_runner" && profile.credential === "OPENROUTER_API_KEY";
|
|
return { ...payload, name: "CEO", title: "CEO", role: "ceo",
|
|
capabilities: "Coordinates company work and delegates tasks to the team.",
|
|
budgetMonthlyCents: CHAT_FEATURE_BUDGET_CENTS,
|
|
adapterConfig: { ...adapterConfig, ...(throughputRouting ? { env: { ...(adapterConfig.env as Record<string, unknown>),
|
|
OPENCODE_CONFIG_CONTENT: JSON.stringify({ provider: { openrouter: { models: {
|
|
"deepseek/deepseek-v4.1-flash-1017": { options: { provider: { sort: "throughput" } } },
|
|
} } } }),
|
|
} } : {}), timeoutSec: profile.id === "runner-opencode" ? CHAT_NATIVE_TIMEOUT_POLICY.savedAgentTimeoutSec : 120, ...(profile.provider === "codex" ? { modelReasoningEffort: "low" } : {}) },
|
|
runtimeConfig: { ...(payload.runtimeConfig as Record<string, unknown>),
|
|
heartbeat: { enabled: false, maxConcurrentRuns: 1, maxDailyRuns: 3, maxDailyCostCents: CHAT_FEATURE_BUDGET_CENTS } },
|
|
};
|
|
} };
|
|
}
|
|
export const chatFeatureTasks: readonly RunnerTaskFixture[] = CHAT_FEATURE_CASES.map((id, index) => ({
|
|
id, label: id === "create-backlog" ? "Save a task for Morgan without starting it" : "Delegate an announcement and return its saved result",
|
|
groups: ["chat"], flow: "agent_chat", workMode: "standard",
|
|
expectedRunCount: id === "create-backlog" ? 1 : 3,
|
|
automaticRetryPolicy: "single_attempt", turnTimeoutMs: 120_000,
|
|
attemptTimeoutMs: { local: 420_000, daytona: 420_000 },
|
|
expectedTerminalState: { issue: "in_review", run: "succeeded" },
|
|
buildTitle: nonce => `Chat feature ${id} ${nonce}`,
|
|
buildPrompt: () => chatFeatureScenarios[index]!.query,
|
|
buildVisibleMarker: nonce => `ACORN${nonce.replace(/[^a-zA-Z0-9]/g, "")}`,
|
|
buildMatchers: () => [{ kind: "issue_status", expected: "in_review" }],
|
|
}));
|
|
|
|
|
|
/** Only executable domain cases become selectable cells. Missing cases remain in the frozen denominator. */
|
|
export const chatCatalogTasks: readonly RunnerTaskFixture[] = chatCatalogCases.map(implementation => {
|
|
const scenario = chatCatalogScenarios.find(row => row.id === implementation.scenarioId)!;
|
|
return {
|
|
requiredCredentials: implementation.requiredCredentials,
|
|
id: scenario.id, label: scenario.id, groups: ["chat"], flow: "agent_chat", workMode: "standard",
|
|
expectedRunCount: implementation.expectedRunCount, minimumExpectedRunCount: implementation.minimumRunCount,
|
|
...(implementation.optionalCompanySecretBindingApproval===true?{expectedHumanMessageCount:implementation.expectedHumanMessageCount,optionalCompanySecretBindingApproval:true as const}:{}),
|
|
automaticRetryPolicy: "single_attempt", turnTimeoutMs: CHAT_NATIVE_TIMEOUT_POLICY.effectiveNativeTurnTimeoutMs,
|
|
attemptTimeoutMs: { local: implementation.attemptTimeoutMs, daytona: implementation.attemptTimeoutMs },
|
|
expectedTerminalState: { issue: "in_review", run: "succeeded" },
|
|
buildTitle: nonce => `Chat catalog ${scenario.id} ${nonce}`,
|
|
buildPrompt: () => scenario.query,
|
|
buildVisibleMarker: nonce => `CATALOG${nonce.replace(/[^a-zA-Z0-9]/g, "")}`,
|
|
buildMatchers: () => [{ kind: "issue_status", expected: "in_review" }],
|
|
};
|
|
});
|