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>
65 lines
5.9 KiB
JavaScript
65 lines
5.9 KiB
JavaScript
import assert from "node:assert/strict";
|
|
import { readFileSync } from "node:fs";
|
|
import { EventEmitter } from "node:events";
|
|
import vm from "node:vm";
|
|
import { transformSync } from "esbuild";
|
|
import { it } from "vitest";
|
|
|
|
it("stages company authority only across the actual completed supervisor restart boundary", async () => {
|
|
const source = readFileSync(new URL('./server.ts', import.meta.url), 'utf8');
|
|
const extract=(start,end)=>{const a=source.indexOf(start),b=source.indexOf(end,a);assert.ok(a>=0&&b>a);const part=source.slice(a,b);assert.ok(part.trimEnd().endsWith('}'));return part;};
|
|
const functions=extract('async function startServer()', '\nfunction delay(')+'\n'+extract('async function restartServer(', '\nfor (const signal of');
|
|
const controls=[];
|
|
const grantA={schema:'paperclip.acpx-candidate-admission.v1',companyId:'11111111-1111-4111-8111-111111111111',agent:'kimi',model:'kimi-for-coding',platform:'darwin',architecture:'arm64',expiresAt:Date.now()+60000};
|
|
const grantB={...grantA,companyId:'22222222-2222-4222-8222-222222222222'};
|
|
function factory(code){
|
|
const events=[],spawns=[];let pid=100,stopFails=false,shuttingDown=false;
|
|
const stop=async c=>{events.push('stop');if(stopFails)throw new Error('controlled stop failure');c.exitCode=0;events.push('stopped');};
|
|
stop.watch=async()=>{};stop.forRestart=stop;stop.crashForRestart=stop;
|
|
const ctx=vm.createContext({Map,Set,JSON,Date,Error,Promise,
|
|
process:{execPath:'/controlled/pinned-node',env:{PAPERCLIP_ACPX_CANDIDATE_ADMISSION:'ambient-forgery'},stdout:{write(){}},stderr:{write(){}}},
|
|
upgradeArtifacts:undefined,artifactIndex:0,selectedExecutions:['controlled-kimi'],
|
|
CHAT_INSTALL_PLAN:'CONTROLLED_CHAT_INSTALL_PLAN',temporaryRoot:'/controlled/tmp',invitationSourceGeneration:undefined,
|
|
invitationSourceProofPath:()=>'/controlled/tmp/proof',rm:async()=>{},executionIds:[],invitationAttemptId:'controlled',
|
|
publishInvitationSourceProof:async()=>undefined,
|
|
verifyInstalledCli:async()=>null,installedReleaseLaunch:()=>{throw Error('unexpected installed path');},
|
|
definedServerEnvironment:{PATH:'/controlled/bin'},repositoryRoot:'/controlled/repo',paperclipCli:'/controlled/cli',runnerE2EServerDetached:true,
|
|
runnerE2ETypeScriptProcessArgs:()=>['controlled-cli','onboard','--yes','--run'],
|
|
shutdownRequested:()=>shuttingDown,shutdownSignal:null,expectedStops:new Set(),childErrors:new Map(),servingArtifacts:new Map(),
|
|
kimiCandidateAdmission:undefined,activeRestartRequestId:null,child:null,log:{write(){}},appendLog(){},
|
|
stopServer:stop,waitForHealthToStop:async()=>events.push('health-stopped'),waitForHealth:async()=>events.push('healthy'),writeRestartAck:async()=>events.push('ack'),
|
|
spawn:(executable,args,options)=>{events.push('spawn');const c=new EventEmitter();c.pid=++pid;c.stdout=new EventEmitter();c.stderr=new EventEmitter();c.exitCode=null;c.signalCode=null;spawns.push({executable,args,options});return c;},
|
|
});
|
|
vm.runInContext(transformSync(code,{loader:'ts',target:'es2022',format:'cjs'}).code+'\nglobalThis.control = {startServer,restartServer};',ctx);
|
|
return {ctx,events,spawns,setStopFails:v=>stopFails=v,setShutdown:v=>shuttingDown=v,...ctx.control};
|
|
}
|
|
const old=factory(functions.replace('if (stagedKimiAdmission) kimiCandidateAdmission = stagedKimiAdmission;',''));
|
|
await old.startServer();await old.restartServer({requestId:'counterfactual-missing-assignment',kimiCandidateAdmission:grantA});
|
|
assert.equal(old.spawns.at(-1).options.env.PAPERCLIP_ACPX_CANDIDATE_ADMISSION,undefined);
|
|
controls.push({name:'removed-assignment counterfactual reaches real spawn without staged authority',result:'RED_REPRODUCED',counterfactualOnly:true});
|
|
const current=factory(functions);
|
|
await current.startServer();assert.equal(current.spawns[0].options.env.PAPERCLIP_ACPX_CANDIDATE_ADMISSION,undefined);
|
|
controls.push({name:'initial start ignores ambient authority and has no supervisor-held record',result:'PASS'});
|
|
current.events.length=0;
|
|
await current.restartServer({requestId:'stage-authority',kimiCandidateAdmission:grantA});
|
|
assert.deepEqual(JSON.parse(current.spawns.at(-1).options.env.PAPERCLIP_ACPX_CANDIDATE_ADMISSION),grantA);
|
|
assert.ok(current.events.indexOf('stopped')<current.events.indexOf('spawn'));
|
|
assert.ok(current.events.indexOf('health-stopped')<current.events.indexOf('spawn'));
|
|
assert.ok(current.events.indexOf('healthy')<current.events.indexOf('ack'));
|
|
controls.push({name:'validated staged authority reaches actual spawn after old process/health stop and before ready acknowledgement',result:'PASS'});
|
|
await current.restartServer({requestId:'ordinary-restart'});
|
|
assert.deepEqual(JSON.parse(current.spawns.at(-1).options.env.PAPERCLIP_ACPX_CANDIDATE_ADMISSION),grantA);
|
|
controls.push({name:'ordinary later restart preserves exact held authority',result:'PASS'});
|
|
const before=current.spawns.length;current.setStopFails(true);
|
|
await assert.rejects(current.restartServer({requestId:'stop-failure',kimiCandidateAdmission:grantB}),/controlled stop failure/);
|
|
assert.equal(current.spawns.length,before);assert.deepEqual(current.ctx.kimiCandidateAdmission,grantA);
|
|
controls.push({name:'failed old stop neither changes held authority nor spawns replacement',result:'PASS'});
|
|
current.setStopFails(false);current.setShutdown(true);
|
|
await assert.rejects(current.restartServer({requestId:'shutdown',kimiCandidateAdmission:grantB}),/Wrapper shutdown/);
|
|
assert.equal(current.spawns.length,before);assert.deepEqual(current.ctx.kimiCandidateAdmission,grantA);
|
|
controls.push({name:'shutdown before replacement neither changes held authority nor spawns replacement',result:'PASS'});
|
|
current.setShutdown(false);await current.startServer();await current.restartServer({requestId:'next-company',kimiCandidateAdmission:grantB});
|
|
assert.deepEqual(JSON.parse(current.spawns.at(-1).options.env.PAPERCLIP_ACPX_CANDIDATE_ADMISSION),grantB);
|
|
controls.push({name:'next validated disposable company replaces record only on explicit staging restart',result:'PASS'});
|
|
});
|