| 1 | // End-to-end agent tasks: a real `wrangler dev` Worker, sandboxes in an |
| 2 | // `agent.mjs --sandbox-pool` process (containers don't run under wrangler dev), |
| 3 | // a fake model provider behind the model proxy, and the real git CLI. |
| 4 | // |
| 5 | // pnpm --filter @diffinery/app e2e |
| 6 | // |
| 7 | // Covers model keys (write-only, browser session only, opt-in), a task run end |
| 8 | // to end (sandbox as the agent, the model proxy adding the person's key, the |
| 9 | // push and the pull request as the agent), follow-ups, stopping, and the |
| 10 | // security rules: no token, agent or sandbox can start or steer a task, task |
| 11 | // sandboxes are closed to the sandbox API, and a run's model token stops |
| 12 | // working when the run ends. |
| 13 | import "./fresh-connections.mjs"; |
| 14 | import { spawn, execFileSync } from "node:child_process"; |
| 15 | import { mkdtempSync, writeFileSync, mkdirSync, rmSync, readdirSync, statSync, readFileSync, existsSync } from "node:fs"; |
| 16 | import { tmpdir } from "node:os"; |
| 17 | import { createHash, randomBytes } from "node:crypto"; |
| 18 | import { createRequire } from "node:module"; |
| 19 | import { join, dirname, resolve } from "node:path"; |
| 20 | import { fileURLToPath } from "node:url"; |
| 21 | import { createServer as netServer } from "node:net"; |
| 22 | import { createServer as httpServer } from "node:http"; |
| 23 | import { test } from "node:test"; |
| 24 | import assert from "node:assert/strict"; |
| 25 | import { |
| 26 | APP, |
| 27 | AGENT, |
| 28 | work, |
| 29 | poolRoot, |
| 30 | WS, |
| 31 | RIVAL_SESSION, |
| 32 | ANTHROPIC_KEY, |
| 33 | OPENROUTER_KEY, |
| 34 | base, |
| 35 | dev, |
| 36 | pool, |
| 37 | upstream, |
| 38 | devLog, |
| 39 | poolLog, |
| 40 | token, |
| 41 | modelRequests, |
| 42 | jailToken, |
| 43 | api, |
| 44 | ok, |
| 45 | asPerson, |
| 46 | sleep, |
| 47 | until, |
| 48 | git, |
| 49 | mcp, |
| 50 | poolSandboxes, |
| 51 | } from "./agents-harness.mjs"; |
| 52 | import { freePort } from "./ports.mjs"; |
| 53 | |
| 54 | test("model keys: write-only, browser session only, with an opt-in", async () => { |
| 55 | assert.deepEqual(await ok("GET", "/diffinery/model-keys"), []); |
| 56 | // No token can set one, however wide. |
| 57 | const viaToken = await api("PUT", "/diffinery/model-keys/anthropic", { key: ANTHROPIC_KEY, accept_terms: true }); |
| 58 | assert.equal(viaToken.status, 403, JSON.stringify(viaToken.json)); |
| 59 | assert.match(viaToken.json.message, /browser session/); |
| 60 | // The opt-in is required, and the key's shape is checked. |
| 61 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/anthropic", { key: ANTHROPIC_KEY })).status, 422); |
| 62 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/anthropic", { key: "sk-proj-not-anthropic-1234567890abcdef", accept_terms: true })).status, 422); |
| 63 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/gemini", { key: ANTHROPIC_KEY, accept_terms: true })).status, 404); |
| 64 | const saved = await asPerson("PUT", "/diffinery/model-keys/anthropic", { key: ANTHROPIC_KEY, accept_terms: true }); |
| 65 | assert.equal(saved.status, 200, JSON.stringify(saved.json)); |
| 66 | assert.equal(saved.json.hint, ANTHROPIC_KEY.slice(-4)); |
| 67 | assert.ok(saved.json.opted_in_at); |
| 68 | // Never returned: not by the API, not on the page. |
| 69 | const listed = await ok("GET", "/diffinery/model-keys"); |
| 70 | assert.deepEqual(listed.map((k) => k.provider), ["anthropic"]); |
| 71 | assert.ok(!JSON.stringify(listed).includes(ANTHROPIC_KEY)); |
| 72 | const page = await (await fetch(`${base}/settings/model-access`)).text(); |
| 73 | assert.match(page, /Model access/); |
| 74 | assert.match(page, /Commercial Terms/); |
| 75 | assert.ok(page.includes(`…${ANTHROPIC_KEY.slice(-4)}`)); |
| 76 | assert.ok(!page.includes(ANTHROPIC_KEY)); |
| 77 | assert.equal((await api("DELETE", "/diffinery/model-keys/anthropic")).status, 403); |
| 78 | }); |
| 79 | |
| 80 | test("an agent task runs in a sandbox as the agent, through the model proxy, and opens a pull request", async () => { |
| 81 | const me = await ok("GET", "/user"); |
| 82 | const owner = me.login; |
| 83 | await ok("POST", "/user/repos", { name: "tasks-demo" }); |
| 84 | const src = join(work, "src"); |
| 85 | mkdirSync(src, { recursive: true }); |
| 86 | writeFileSync(join(src, "README.md"), "# tasks\n"); |
| 87 | git(src, ["init", "-q", "-b", "main"]); |
| 88 | git(src, ["add", "-A"]); |
| 89 | git(src, ["commit", "-q", "-m", "initial"]); |
| 90 | git(src, ["remote", "add", "origin", `${base}/${owner}/tasks-demo.git`]); |
| 91 | git(src, ["push", "-q", "origin", "main"]); |
| 92 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 93 | const bot = (await asPerson("POST", "/diffinery/agents", { login: "task-bot", name: "Task Bot" })).json.token.token; |
| 94 | |
| 95 | const tokFile = join(work, "model-token"); |
| 96 | // The custom runtime: call the model through the proxy, show the key variable, save it for later, change a file. |
| 97 | // It runs in a jail (phase 5d): what it can see and reach is shown too. |
| 98 | const command = [ |
| 99 | `curl -sS -X POST "$ANTHROPIC_BASE_URL/v1/messages" -H "x-api-key: $ANTHROPIC_API_KEY" -H "content-type: application/json" -d '{"model":"m","messages":[]}'`, |
| 100 | "echo", |
| 101 | 'echo "key is $ANTHROPIC_API_KEY"', |
| 102 | `printf %s "$ANTHROPIC_API_KEY" > ${tokFile}`, |
| 103 | 'echo "goal: $DIFFINERY_TASK_PROMPT" | tail -1', |
| 104 | 'echo "jail: credential=$(cat ../repo/.git/config 2>/dev/null || echo hidden)"', |
| 105 | 'echo "jail: token=$(env | grep -c dmt_)"', |
| 106 | // Only the model endpoints through the jail's model proxy, however the path is spelled; Diffinery itself is out of reach. |
| 107 | `echo "jail: api=$(curl -s -o /dev/null -w '%{http_code}' --path-as-is "$ANTHROPIC_BASE_URL/../../api/v3/user")"`, |
| 108 | // (The test's Diffinery is on 127.0.0.1, which NO_PROXY sends direct; --noproxy '' sends it through the proxy.) |
| 109 | `echo "jail: via-proxy=$(curl -s -o /dev/null -w '%{http_code}' --noproxy '' ${base}/api/v3/user)"`, |
| 110 | `echo "jail: direct=$(curl -s -o /dev/null -w '%{http_code}' --noproxy '*' --max-time 5 ${base}/api/v3/user)"`, |
| 111 | "echo hello > hello.txt", |
| 112 | ].join("; "); |
| 113 | const goal = "Add a hello file"; |
| 114 | |
| 115 | // Nobody but a person in a browser starts a task: not their token, not the agent's. |
| 116 | for (const tok of [token, bot]) { |
| 117 | const r = await api("POST", R, { goal, runtime: "custom", command, agent: "task-bot" }, tok); |
| 118 | assert.equal(r.status, 403, JSON.stringify(r.json)); |
| 119 | assert.match(r.json.message, /browser session/); |
| 120 | } |
| 121 | // Codex needs an OpenAI key the person doesn't have. |
| 122 | const codex = await asPerson("POST", R, { goal, runtime: "codex", agent: "task-bot" }); |
| 123 | assert.equal(codex.status, 422); |
| 124 | assert.match(codex.json.message, /OpenAI API key/); |
| 125 | assert.equal((await asPerson("POST", R, { goal, runtime: "custom", command, agent: owner })).status, 422, "a person isn't an agent"); |
| 126 | |
| 127 | const started = await asPerson("POST", R, { goal, runtime: "custom", command, agent: "task-bot" }); |
| 128 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 129 | const n = started.json.number; |
| 130 | assert.equal(started.json.state, "running"); |
| 131 | assert.equal(started.json.branch, "agent/1-add-a-hello-file"); |
| 132 | assert.equal(started.json.agent.login, "task-bot"); |
| 133 | assert.equal(started.json.started_by.login, owner); |
| 134 | |
| 135 | const done = await until("the task to finish", async () => { |
| 136 | const t = await ok("GET", `${R}/${n}`); |
| 137 | return t.state !== "running" ? t : null; |
| 138 | }); |
| 139 | const lines = await ok("GET", `${R}/${n}/transcript`); |
| 140 | const text = lines.map((l) => `[${l.stream}] ${l.text}`).join("\n"); |
| 141 | assert.equal(done.state, "ready", text); |
| 142 | assert.ok(done.pull_request, text); |
| 143 | |
| 144 | // The provider got the person's key, never the run's token; the sandbox never saw the key. |
| 145 | assert.equal(modelRequests.length, 1, text); |
| 146 | assert.equal(modelRequests[0].headers["x-api-key"], ANTHROPIC_KEY); |
| 147 | assert.equal(modelRequests[0].url, "/v1/messages"); |
| 148 | assert.ok(!JSON.stringify(modelRequests[0].headers).includes("dmt_")); |
| 149 | assert.match(text, /fake model says hi/); |
| 150 | // The agent had a stand-in for a key: the run's token stayed outside its jail, with the jail's model proxy. |
| 151 | assert.match(text, /key is diffinery-jail/); |
| 152 | assert.equal(readFileSync(tokFile, "utf8"), "diffinery-jail"); |
| 153 | assert.match(text, /\[out\] Add a hello file/, "the prompt reaches the runtime"); |
| 154 | assert.ok(!text.includes(ANTHROPIC_KEY)); |
| 155 | assert.ok(!text.includes("dmt_")); |
| 156 | // Its jail (phase 5d): no sandbox credential, no token, Diffinery reachable only at its model endpoints, no network of its own. |
| 157 | assert.match(text, /jail: credential=hidden/); |
| 158 | assert.match(text, /jail: token=0/); |
| 159 | assert.match(text, /jail: api=403/); |
| 160 | assert.match(text, /jail: via-proxy=403/); |
| 161 | assert.match(text, /jail: direct=000/); |
| 162 | assert.equal(modelRequests.length, 1); |
| 163 | // The proxy doesn't take an account token, or anything outside /v1/. |
| 164 | assert.equal((await fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", headers: { authorization: `Bearer ${token}` } })).status, 401); |
| 165 | assert.equal((await fetch(`${base}/-/model/anthropic/admin`, { headers: { "x-api-key": `dmt_${"x".repeat(12)}_y` } })).status, 403); |
| 166 | |
| 167 | // The pull request is the agent's, from the task's branch, with the work. |
| 168 | const pr = await ok("GET", `/repos/${owner}/tasks-demo/pulls/${done.pull_request.number}`); |
| 169 | assert.equal(pr.user.login, "task-bot"); |
| 170 | assert.equal(pr.head.ref, "agent/1-add-a-hello-file"); |
| 171 | assert.equal(pr.intent, goal); |
| 172 | const files = await ok("GET", `/repos/${owner}/tasks-demo/pulls/${pr.number}/files`); |
| 173 | assert.deepEqual(files.map((f) => f.filename), ["hello.txt"]); |
| 174 | const commit = await ok("GET", `/repos/${owner}/tasks-demo/commits/${pr.head.sha}`); |
| 175 | assert.equal(commit.commit.author.name, "task-bot"); |
| 176 | |
| 177 | // Read over MCP and on Today; steering is not on MCP. |
| 178 | const viaMcp = await mcp("agent_task", { owner, repo: "tasks-demo", number: n }); |
| 179 | assert.equal(viaMcp.isError, false, viaMcp.content[0].text); |
| 180 | assert.equal(viaMcp.structuredContent.state, "ready"); |
| 181 | assert.ok(viaMcp.structuredContent.transcript.length > 0); |
| 182 | const mine = await mcp("agent_tasks", {}); |
| 183 | assert.deepEqual(mine.structuredContent.result.map((t) => t.number), [n]); |
| 184 | const today = await ok("GET", "/diffinery/today"); |
| 185 | assert.equal(today.tasks[0].number, n); |
| 186 | assert.equal(today.tasks[0].state, "ready"); |
| 187 | |
| 188 | // Pages. |
| 189 | for (const path of ["tasks", `tasks/${n}`, "tasks/new"]) { |
| 190 | const res = await fetch(`${base}/${owner}/tasks-demo/${path}`); |
| 191 | assert.equal(res.status, 200, path); |
| 192 | } |
| 193 | const taskPage = await (await fetch(`${base}/${owner}/tasks-demo/tasks/${n}`)).text(); |
| 194 | assert.match(taskPage, /Add a hello file/); |
| 195 | assert.match(taskPage, /fake model says hi/); |
| 196 | |
| 197 | // Usage (phase 5a): counted by the model proxy from the provider's answer, on the task, its page and the person's 30 days. |
| 198 | const counted = await until("the run's usage", async () => { |
| 199 | const t = await ok("GET", `${R}/${n}`); |
| 200 | return t.runs[0].usage?.input_tokens === 21 ? t : null; |
| 201 | }); |
| 202 | assert.deepEqual(counted.runs[0].usage, { source: "proxy", requests: 1, input_tokens: 21, output_tokens: 4, cache_read_tokens: 300, cache_write_tokens: 0, cost_usd: null }); |
| 203 | assert.deepEqual(counted.usage, { requests: 1, input_tokens: 21, output_tokens: 4, cache_read_tokens: 300, cache_write_tokens: 0, cost_usd: null }); |
| 204 | assert.equal((await mcp("agent_task", { owner, repo: "tasks-demo", number: n })).structuredContent.usage.input_tokens, 21); |
| 205 | const usagePage = await (await fetch(`${base}/${owner}/tasks-demo/tasks/${n}`)).text(); |
| 206 | assert.match(usagePage, /<h2>Usage<\/h2>/); |
| 207 | assert.match(usagePage, /Diffinery<\/td>/); |
| 208 | const report = await until("the usage report", async () => { |
| 209 | const r = await ok("GET", "/diffinery/model-usage"); |
| 210 | return r.repositories.find((x) => x.repository === `${owner}/tasks-demo` && x.kind === "task" && x.input_tokens >= 21) ? r : null; |
| 211 | }); |
| 212 | assert.equal(report.days, 30); |
| 213 | assert.deepEqual(report.repositories[0], { repository: `${owner}/tasks-demo`, kind: "task", runs: 1, reported_by_cli_runs: 0, requests: 1, input_tokens: 21, output_tokens: 4, cache_read_tokens: 300, cache_write_tokens: 0, cost_usd: null }); |
| 214 | assert.equal((await ok("GET", "/diffinery/model-usage?days=500")).days, 90); |
| 215 | // A repository is named as it was when the run ended, not by a later name (Security, 5a). |
| 216 | await ok("PATCH", `/repos/${owner}/tasks-demo`, { name: "tasks-demo-renamed" }); |
| 217 | assert.deepEqual((await ok("GET", "/diffinery/model-usage")).repositories.map((x) => x.repository), [`${owner}/tasks-demo`]); |
| 218 | await ok("PATCH", `/repos/${owner}/tasks-demo-renamed`, { name: "tasks-demo" }); |
| 219 | const settingsPage = await (await fetch(`${base}/settings/model-access`)).text(); |
| 220 | assert.match(settingsPage, /Usage in the last 30 days/); |
| 221 | assert.ok(settingsPage.includes(`>${owner}/tasks-demo</a>`)); |
| 222 | // Someone else's report is their own: the rival pays for nothing here. |
| 223 | const rivalUsage = await (await fetch(`${base}/api/v3/diffinery/model-usage`, { headers: { cookie: `__Host-dfy_session=${RIVAL_SESSION}` } })).json(); |
| 224 | assert.deepEqual(rivalUsage.repositories, []); |
| 225 | |
| 226 | // Follow-ups and retries: the person, in a browser; never a token. |
| 227 | assert.equal((await api("POST", `${R}/${n}/messages`, { body: "again" })).status, 403); |
| 228 | assert.equal((await api("POST", `${R}/${n}/retry`, {})).status, 403); |
| 229 | assert.equal((await api("POST", `${R}/${n}/messages`, { body: "again" }, bot)).status, 403); |
| 230 | const followUp = await asPerson("POST", `${R}/${n}/messages`, { body: "Say hello again" }); |
| 231 | assert.equal(followUp.status, 201, JSON.stringify(followUp.json)); |
| 232 | const second = await until("the follow-up run", async () => { |
| 233 | const t = await ok("GET", `${R}/${n}`); |
| 234 | return t.state !== "running" ? t : null; |
| 235 | }); |
| 236 | assert.equal(second.state, "ready"); |
| 237 | assert.deepEqual(second.runs.map((r) => [r.kind, r.state]), [ |
| 238 | ["start", "done"], |
| 239 | ["message", "done"], |
| 240 | ]); |
| 241 | const text2 = (await ok("GET", `${R}/${n}/transcript?after=${lines.at(-1).seq}`)).map((l) => l.text).join("\n"); |
| 242 | assert.match(text2, /Say hello again/); |
| 243 | assert.match(text2, /No changes to push/); |
| 244 | assert.match(text2, /Starting a sandbox on agent\/1-add-a-hello-file/); |
| 245 | }); |
| 246 | |
| 247 | test("task sandboxes are closed to the sandbox API, and stopping ends the run's access", async () => { |
| 248 | const me = await ok("GET", "/user"); |
| 249 | const owner = me.login; |
| 250 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 251 | const bot = (await asPerson("POST", "/diffinery/agents/task-bot/tokens", { name: "second" })).json.token; |
| 252 | const tokFile = join(work, "model-token-2"); |
| 253 | // An OpenRouter key too: its own shape, its own opt-in. |
| 254 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/openrouter", { key: ANTHROPIC_KEY, accept_terms: true })).status, 422); |
| 255 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/openrouter", { key: OPENROUTER_KEY })).status, 422); |
| 256 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/openrouter", { key: OPENROUTER_KEY, accept_terms: true })).status, 200); |
| 257 | const before_ = new Set(poolSandboxes()); |
| 258 | const started = await asPerson("POST", R, { goal: "Wait a while", runtime: "custom", command: `printf %s "$ANTHROPIC_API_KEY" > ${tokFile}; sleep 120`, agent: "task-bot" }); |
| 259 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 260 | const n = started.json.number; |
| 261 | const id = await until("the task's sandbox", async () => poolSandboxes().find((x) => !before_.has(x)) ?? null); |
| 262 | const runToken = await jailToken(tokFile); |
| 263 | |
| 264 | // While it runs, its token works through the proxy. |
| 265 | const live = await fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", headers: { "x-api-key": runToken, "content-type": "application/json" }, body: "{}" }); |
| 266 | assert.equal(live.status, 200); |
| 267 | |
| 268 | // Only the model endpoints: not the rest of the person's provider account. |
| 269 | for (const [method, path] of [ |
| 270 | ["GET", "/v1/files"], |
| 271 | ["POST", "/v1/messages/batches"], |
| 272 | ["DELETE", "/v1/files/file_1"], |
| 273 | ]) { |
| 274 | const seen = modelRequests.length; |
| 275 | const r = await fetch(`${base}/-/model/anthropic${path}`, { method, headers: { "x-api-key": runToken, "content-type": "application/json" }, body: method === "POST" ? "{}" : undefined }); |
| 276 | assert.equal(r.status, 403, `${method} ${path}`); |
| 277 | assert.match((await r.json()).error.message, new RegExp(path.replace(/\//g, "\\/"))); |
| 278 | assert.equal(modelRequests.length, seen, `${method} ${path} must not reach the provider`); |
| 279 | } |
| 280 | |
| 281 | // OpenRouter: the run's token works there too (the custom runtime gets every provider the person has a key for), |
| 282 | // with the person's OpenRouter key and Diffinery's attribution added, and only its model endpoints. |
| 283 | const seenOr = modelRequests.length; |
| 284 | const orLive = await fetch(`${base}/-/model/openrouter/v1/chat/completions`, { method: "POST", headers: { authorization: `Bearer ${runToken}`, "content-type": "application/json", "x-title": "spoofed", "http-referer": "https://spoofed.example", referer: "https://spoofed.example", "x-openrouter-categories": "spoofed" }, body: '{"model":"anthropic/claude-sonnet-4.5","messages":[]}' }); |
| 285 | assert.equal(orLive.status, 200); |
| 286 | const orReq = modelRequests[seenOr]; |
| 287 | assert.equal(orReq.url, "/v1/chat/completions"); |
| 288 | assert.equal(orReq.headers.authorization, `Bearer ${OPENROUTER_KEY}`); |
| 289 | assert.equal(orReq.headers["x-api-key"], undefined); |
| 290 | assert.equal(orReq.headers["x-title"], "Diffinery"); |
| 291 | assert.equal(orReq.headers["http-referer"], base); |
| 292 | // Only Diffinery names itself to OpenRouter: the sandbox's attribution headers never reach it. |
| 293 | assert.equal(orReq.headers.referer, undefined); |
| 294 | assert.equal(orReq.headers["x-openrouter-categories"], undefined); |
| 295 | // Claude Code's way in: the Messages endpoint, with the token as a bearer token. |
| 296 | assert.equal((await fetch(`${base}/-/model/openrouter/v1/messages`, { method: "POST", headers: { authorization: `Bearer ${runToken}`, "content-type": "application/json" }, body: "{}" })).status, 200); |
| 297 | for (const [method, path] of [ |
| 298 | ["GET", "/v1/key"], |
| 299 | ["GET", "/v1/credits"], |
| 300 | ["GET", "/v1/keys"], |
| 301 | ["POST", "/v1/keys"], |
| 302 | ["GET", "/v1/generation"], |
| 303 | ["GET", "/v1/activity"], |
| 304 | ["GET", "/v1/models/user"], |
| 305 | ]) { |
| 306 | const seen = modelRequests.length; |
| 307 | const r = await fetch(`${base}/-/model/openrouter${path}`, { method, headers: { authorization: `Bearer ${runToken}`, "content-type": "application/json" }, body: method === "POST" ? "{}" : undefined }); |
| 308 | assert.equal(r.status, 403, `openrouter ${method} ${path}`); |
| 309 | assert.equal(modelRequests.length, seen, `openrouter ${method} ${path} must not reach the provider`); |
| 310 | } |
| 311 | |
| 312 | // No public sandbox call reaches it: the person (a repository admin here), their token, the agent. |
| 313 | for (const [who, tok] of [ |
| 314 | ["the person in a browser", null], |
| 315 | ["the person's admin token", token], |
| 316 | ["the agent's token", bot], |
| 317 | ]) { |
| 318 | for (const action of ["exec", "read", "write", "close"]) { |
| 319 | const r = await api("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes/${id}/${action}`, { cmd: "env", path: "README.md", content: "" }, tok); |
| 320 | assert.equal(r.status, 404, `${who} ${action}: ${JSON.stringify(r.json)}`); |
| 321 | } |
| 322 | } |
| 323 | for (const tool of ["sandbox_exec", "sandbox_read_file", "sandbox_write_file", "sandbox_close"]) { |
| 324 | const r = await mcp(tool, { owner, repo: "tasks-demo", sandbox_id: id, command: "env", path: "README.md", content: "" }, bot); |
| 325 | assert.equal(r.isError, true, tool); |
| 326 | assert.match(r.content[0].text, /sandbox not found/, tool); |
| 327 | } |
| 328 | |
| 329 | // A plain sandbox's own credential can't start or steer a task either. |
| 330 | const plain = await ok("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes`, {}, bot); |
| 331 | const cfgPath = join(poolRoot, plain.id, "repo", ".git", "config"); |
| 332 | const config = await until("the plain sandbox's git credential", async () => (existsSync(cfgPath) && /extraheader/i.test(readFileSync(cfgPath, "utf8")) ? readFileSync(cfgPath, "utf8") : null)); |
| 333 | const basic = /extraheader = Authorization: Basic (\S+)/i.exec(config); |
| 334 | assert.ok(basic, config); |
| 335 | const dst = Buffer.from(basic[1], "base64").toString().split(":").pop(); |
| 336 | assert.match(dst, /^dst_/); |
| 337 | assert.equal((await api("POST", R, { goal: "x", runtime: "custom", command: "true", agent: "task-bot" }, dst)).status, 403); |
| 338 | assert.equal((await api("POST", `${R}/${n}/messages`, { body: "x" }, dst)).status, 403); |
| 339 | await ok("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes/${plain.id}/close`, {}, bot); |
| 340 | |
| 341 | // One account keeps at most 5 sandboxes open in a repository; closing one makes room. |
| 342 | const open = []; |
| 343 | for (let i = 0; i < 5; i++) open.push((await ok("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes`, {}, bot)).id); |
| 344 | const sixth = await api("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes`, {}, bot); |
| 345 | assert.equal(sixth.status, 429); |
| 346 | assert.match(sixth.json.message, /5 open sandboxes in this repository/); |
| 347 | await ok("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes/${open.pop()}/close`, {}, bot); |
| 348 | open.push((await ok("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes`, {}, bot)).id); |
| 349 | for (const id of open) await ok("POST", `/repos/${owner}/tasks-demo/diffinery/sandboxes/${id}/close`, {}, bot); |
| 350 | |
| 351 | // The agent can stop its own task; the run's token stops working. |
| 352 | const stopped = await ok("POST", `${R}/${n}/stop`, {}, bot); |
| 353 | assert.equal(stopped.state, "stopped"); |
| 354 | const after_ = await fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", headers: { "x-api-key": runToken, "content-type": "application/json" }, body: "{}" }); |
| 355 | assert.equal(after_.status, 401); |
| 356 | const t = await ok("GET", `${R}/${n}`); |
| 357 | assert.equal(t.runs[0].state, "stopped"); |
| 358 | assert.match((await ok("GET", `${R}/${n}/transcript`)).map((l) => l.text).join("\n"), /Stopped by task-bot/); |
| 359 | |
| 360 | // Claude Code and Codex can run through OpenRouter with a model of the person's choice; only combinations that work. |
| 361 | const form = await (await fetch(`${base}/${owner}/tasks-demo/tasks/new`)).text(); |
| 362 | assert.match(form, /claude-code\+openrouter","label":"Claude Code via OpenRouter/); |
| 363 | assert.match(form, /codex\+openrouter","label":"Codex via OpenRouter/); |
| 364 | for (const bad of [ |
| 365 | { runtime: "codex", provider: "anthropic" }, |
| 366 | { runtime: "custom", command: "true", provider: "openrouter" }, |
| 367 | { runtime: "claude-code", provider: "openrouter", model: "not a model" }, |
| 368 | { runtime: "claude-code", model: "anthropic/claude-sonnet-4.5" }, |
| 369 | ]) assert.equal((await asPerson("POST", R, { goal: "x", agent: "task-bot", ...bad })).status, 422, JSON.stringify(bad)); |
| 370 | const viaOr = await asPerson("POST", R, { goal: "Say hi", agent: "task-bot", runtime: "claude-code", provider: "openrouter", model: "anthropic/claude-sonnet-4.5" }); |
| 371 | assert.equal(viaOr.status, 201, JSON.stringify(viaOr.json)); |
| 372 | assert.equal(viaOr.json.runtime, "claude-code"); |
| 373 | assert.equal(viaOr.json.provider, "openrouter"); |
| 374 | assert.equal(viaOr.json.model, "anthropic/claude-sonnet-4.5"); |
| 375 | assert.equal((await ok("GET", `${R}/${viaOr.json.number}`)).provider, "openrouter"); |
| 376 | await ok("POST", `${R}/${viaOr.json.number}/stop`, {}, bot); |
| 377 | }); |
| 378 | |
| 379 | test("a repository is a project: settings, memory, conversation, and a queue of tasks", async () => { |
| 380 | const me = await ok("GET", "/user"); |
| 381 | const owner = me.login; |
| 382 | const P = `/repos/${owner}/tasks-demo/diffinery/project`; |
| 383 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 384 | |
| 385 | // Settings: an admin in a browser; never a token. |
| 386 | assert.equal((await api("PUT", `${P}/settings`, { max_running: 1 })).status, 403); |
| 387 | for (const bad of [0, 11, 1.5, "x"]) assert.equal((await asPerson("PUT", `${P}/settings`, { max_running: bad })).status, 422, String(bad)); |
| 388 | assert.deepEqual((await asPerson("PUT", `${P}/settings`, { max_running: 1, instructions: "Always say PROJECT-RULE-7." })).json, { max_running: 1, instructions: "Always say PROJECT-RULE-7.", plan_first: "off" }); |
| 389 | |
| 390 | // Memory: every save is a version, and a restore is one more. |
| 391 | assert.equal((await api("PUT", `${P}/memory`, { body: "x" })).status, 403); |
| 392 | assert.equal((await asPerson("PUT", `${P}/memory`, { body: "MEMORY-NOTE-1: docs live in docs/." })).json.version, 1); |
| 393 | assert.equal((await asPerson("PUT", `${P}/memory`, { body: "MEMORY-NOTE-2: something wrong." })).json.version, 2); |
| 394 | assert.equal((await asPerson("PUT", `${P}/memory`, { restore: 99 })).status, 404); |
| 395 | const restored = await asPerson("PUT", `${P}/memory`, { restore: 1 }); |
| 396 | assert.equal(restored.json.version, 3); |
| 397 | assert.equal(restored.json.body, "MEMORY-NOTE-1: docs live in docs/."); |
| 398 | const memory = await ok("GET", `${P}/memory`); |
| 399 | assert.deepEqual(memory.versions.map((v) => v.version), [3, 2, 1]); |
| 400 | assert.equal(memory.updated_by.login, owner); |
| 401 | |
| 402 | // The conversation: people post in a browser. |
| 403 | assert.equal((await api("POST", `${P}/posts`, { body: "hi" })).status, 403); |
| 404 | assert.equal((await asPerson("POST", `${P}/posts`, { body: " " })).status, 422); |
| 405 | const idea = await asPerson("POST", `${P}/posts`, { body: "Idea: write a queued file" }); |
| 406 | assert.equal(idea.status, 201, JSON.stringify(idea.json)); |
| 407 | |
| 408 | // One at a time: the second task waits, then starts when the first ends. Both get the instructions and memory. |
| 409 | const command = (file, wait) => `printf '%s\\n' "$DIFFINERY_TASK_PROMPT" | grep -E 'PROJECT-RULE-7|MEMORY-NOTE|--- (instructions|memory) ---'; sleep ${wait}; echo ${file} > ${file}.txt`; |
| 410 | const a = await asPerson("POST", R, { goal: "Write the first file", runtime: "custom", command: command("first", 8), agent: "task-bot" }); |
| 411 | assert.equal(a.status, 201, JSON.stringify(a.json)); |
| 412 | const b = await asPerson("POST", R, { goal: "Write the queued file", runtime: "custom", command: command("queued", 0), agent: "task-bot" }); |
| 413 | assert.equal(b.status, 201, JSON.stringify(b.json)); |
| 414 | assert.equal(a.json.state, "running"); |
| 415 | assert.equal(b.json.state, "queued"); |
| 416 | const board = await ok("GET", P); |
| 417 | assert.deepEqual(board.board.working.map((t) => t.number), [a.json.number]); |
| 418 | assert.deepEqual(board.board.queued.map((t) => t.number), [b.json.number]); |
| 419 | assert.equal(board.settings.max_running, 1); |
| 420 | assert.equal(board.memory.version, 3); |
| 421 | const queuedLines = (await ok("GET", `${R}/${b.json.number}/transcript`)).map((l) => l.text).join("\n"); |
| 422 | assert.match(queuedLines, /Queued: 1 task is already running/); |
| 423 | const listed = (await ok("GET", `${R}?state=queued`)).map((t) => t.number); |
| 424 | assert.deepEqual(listed, [b.json.number]); |
| 425 | |
| 426 | const first = await until("the first task", async () => { |
| 427 | const t = await ok("GET", `${R}/${a.json.number}`); |
| 428 | return t.state !== "running" ? t : null; |
| 429 | }); |
| 430 | assert.equal(first.state, "ready"); |
| 431 | const second = await until("the queued task", async () => { |
| 432 | const t = await ok("GET", `${R}/${b.json.number}`); |
| 433 | return t.state === "ready" || t.state === "failed" ? t : null; |
| 434 | }); |
| 435 | const text = (await ok("GET", `${R}/${b.json.number}/transcript`)).map((l) => l.text).join("\n"); |
| 436 | assert.equal(second.state, "ready", text); |
| 437 | assert.ok(second.pull_request, text); |
| 438 | // The queued run started only after the first one ended. |
| 439 | assert.ok(Date.parse(second.runs[0].started_at) >= Date.parse(first.runs[0].ended_at), JSON.stringify([first.runs[0], second.runs[0]])); |
| 440 | for (const want of ["--- instructions ---", "PROJECT-RULE-7", "--- memory ---", "MEMORY-NOTE-1"]) assert.ok(text.includes(want), `${want}\n${text}`); |
| 441 | assert.ok(!text.includes("MEMORY-NOTE-2"), "the restored memory, not the replaced one"); |
| 442 | |
| 443 | // Tasks report to the conversation. |
| 444 | const posts = await ok("GET", `${P}/posts`); |
| 445 | const bodies = posts.map((p) => p.body); |
| 446 | assert.ok(bodies.includes("Idea: write a queued file")); |
| 447 | assert.ok(bodies.includes(`${owner} started task #${b.json.number}: Write the queued file`), bodies.join("\n")); |
| 448 | assert.ok(bodies.includes(`Task #${b.json.number} opened pull request #${second.pull_request.number}.`), bodies.join("\n")); |
| 449 | const later = await ok("GET", `${P}/posts?after=${idea.json.id}`); |
| 450 | assert.ok(later.every((p) => p.id > idea.json.id && p.kind === "event")); |
| 451 | |
| 452 | // Over MCP. |
| 453 | const viaMcp = await mcp("project", { owner, repo: "tasks-demo" }); |
| 454 | assert.equal(viaMcp.isError, false, viaMcp.content[0].text); |
| 455 | assert.equal(viaMcp.structuredContent.settings.max_running, 1); |
| 456 | const postsMcp = await mcp("project_posts", { owner, repo: "tasks-demo" }); |
| 457 | assert.equal(postsMcp.isError, false, postsMcp.content[0].text); |
| 458 | assert.ok(JSON.stringify(postsMcp.structuredContent).includes("Idea: write a queued file")); |
| 459 | |
| 460 | // Pages. |
| 461 | const projectPage = await (await fetch(`${base}/${owner}/tasks-demo/project`)).text(); |
| 462 | assert.match(projectPage, /Ready for review/); |
| 463 | assert.match(projectPage, /Write the queued file/); |
| 464 | assert.match(projectPage, /Idea: write a queued file/); |
| 465 | assert.match(projectPage, /MEMORY-NOTE-1/); |
| 466 | assert.match(projectPage, /href="\/[^"]+\/tasks-demo\/project" class="on"/); |
| 467 | const fromPost = await (await fetch(`${base}/${owner}/tasks-demo/tasks/new?post=${idea.json.id}`)).text(); |
| 468 | assert.match(fromPost, /Idea: write a queued file<\/textarea>/); |
| 469 | const memoryPage = await (await fetch(`${base}/${owner}/tasks-demo/project/memory?version=2`)).text(); |
| 470 | assert.match(memoryPage, /Restore this version/); |
| 471 | assert.match(memoryPage, /<tr class="del"><td class="ln">-<\/td><td class="code">MEMORY-NOTE-1/); |
| 472 | assert.match(memoryPage, /<tr class="add"><td class="ln">\+<\/td><td class="code">MEMORY-NOTE-2/); |
| 473 | const settingsPage = await (await fetch(`${base}/${owner}/tasks-demo/settings/project`)).text(); |
| 474 | assert.match(settingsPage, /name="max_running"[^>]*value="1"/); |
| 475 | assert.match(settingsPage, /Always say PROJECT-RULE-7\./); |
| 476 | |
| 477 | // The web forms work too, and raising the limit is what later tests need. |
| 478 | const form = async (path, fields) => { |
| 479 | const res = await fetch(`${base}/${owner}/tasks-demo/settings/project`); |
| 480 | const csrf = /name="_csrf" value="([^"]+)"/.exec(await res.text())?.[1]; |
| 481 | const cookie = /dfy_csrf=([^;]+)/.exec(res.headers.get("set-cookie") ?? "")?.[1]; |
| 482 | return fetch(`${base}/${owner}/tasks-demo/${path}`, { method: "POST", redirect: "manual", headers: { cookie: `dfy_csrf=${cookie}`, "sec-fetch-site": "same-origin" }, body: new URLSearchParams({ _csrf: csrf ?? "", ...fields }) }); |
| 483 | }; |
| 484 | assert.equal((await form("settings/project", { max_running: "0", instructions: "" })).status, 422); |
| 485 | assert.equal((await form("settings/project", { max_running: "3", instructions: "", plan_first: "sometimes" })).status, 422); |
| 486 | assert.equal((await form("settings/project", { max_running: "3", instructions: "", plan_first: "default" })).status, 303); |
| 487 | assert.deepEqual((await ok("GET", P)).settings, { max_running: 3, instructions: "", plan_first: "default" }); |
| 488 | assert.equal((await form("settings/project", { max_running: "3", instructions: "", plan_first: "off" })).status, 303); |
| 489 | assert.equal((await form("project/posts", { body: "Posted from the page" })).status, 303); |
| 490 | assert.ok((await ok("GET", `${P}/posts`)).some((p) => p.body === "Posted from the page")); |
| 491 | assert.equal((await form("project/memory/2/restore", {})).status, 303); |
| 492 | assert.equal((await ok("GET", `${P}/memory`)).body, "MEMORY-NOTE-2: something wrong."); |
| 493 | }); |
| 494 | |
| 495 | test("a coordinator turns posts into tasks, within its caps, and acts only through a validated block", async () => { |
| 496 | const me = await ok("GET", "/user"); |
| 497 | const owner = me.login; |
| 498 | const P = `/repos/${owner}/tasks-demo/diffinery/project`; |
| 499 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 500 | // The coordinator and the tasks it starts run this script (the custom runtime), from the default branch. |
| 501 | const src = join(work, "src"); |
| 502 | writeFileSync( |
| 503 | join(src, "coord.sh"), |
| 504 | `p="$DIFFINERY_TASK_PROMPT" |
| 505 | case "$p" in |
| 506 | *"You are the project coordinator"*) |
| 507 | printf '%s\\n' "$p" | grep -E '^--- (post #|memory ---|instructions ---)' || true |
| 508 | post=$(printf '%s\\n' "$p" | sed -n '/^--- post #/,/^--- end of post ---/p') |
| 509 | case "$post" in |
| 510 | *BAD-BATCH*) printf '%s\\n' 'Checking.' '\`\`\`json' '{"reply": "Stopping it.", "stop": [999]}' '\`\`\`' ;; |
| 511 | *MSG-HAND*) printf '%s\\n' '\`\`\`json' '{"reply": "Nudging it.", "messages": [{"task": 1, "text": "Spend more."}]}' '\`\`\`' ;; |
| 512 | *MSG-OWN*) |
| 513 | n=$(printf '%s\\n' "$p" | sed -n 's/^- #\\([0-9]*\\) \\[[a-z]*\\] Small part .*started by you: you may message.*/\\1/p' | head -1) |
| 514 | printf '%s\\n' '\`\`\`json' "{\\"reply\\": \\"Following up on #$n.\\", \\"messages\\": [{\\"task\\": $n, \\"text\\": \\"Check it again.\\"}]}" '\`\`\`' ;; |
| 515 | *PLAN-THREE*) printf '%s\\n' '\`\`\`json' '{"reply": "Splitting it in three.", "tasks": [{"title": "Small part", "goal": "COORD-SMALL: write the small part."}, {"title": "Big part", "goal": "COORD-BIG: write the big part.", "after": [0]}, {"title": "Never part", "goal": "COORD-NEVER: never runs.", "after": [1]}], "memory": "MEMORY-BY-COORD: parts are split in three."}' '\`\`\`' ;; |
| 516 | *) printf '%s\\n' '\`\`\`json' '{"reply": "Noted, nothing to do."}' '\`\`\`' ;; |
| 517 | esac ;; |
| 518 | *COORD-SMALL*) |
| 519 | curl -s -o /dev/null -w "model %{http_code}\\n" -X POST "$ANTHROPIC_BASE_URL/v1/messages" -H "x-api-key: $ANTHROPIC_API_KEY" -H "content-type: application/json" -d '{"model":"m","max_tokens":5,"messages":[]}' |
| 520 | echo small > small.txt ;; |
| 521 | *COORD-BIG*) |
| 522 | curl -s -o /dev/null -w "model %{http_code}\\n" -X POST "$ANTHROPIC_BASE_URL/v1/messages" -H "x-api-key: $ANTHROPIC_API_KEY" -H "content-type: application/json" -d '{"model":"m","max_tokens":200000,"messages":[]}' |
| 523 | echo big > big.txt ;; |
| 524 | *) echo other > other.txt ;; |
| 525 | esac |
| 526 | `, |
| 527 | ); |
| 528 | git(src, ["pull", "-q", "--rebase", "origin", "main"]); |
| 529 | git(src, ["add", "coord.sh"]); |
| 530 | git(src, ["commit", "-q", "-m", "coordinator script"]); |
| 531 | git(src, ["push", "-q", "origin", "main"]); |
| 532 | |
| 533 | // Your coordinator: yours alone, set in a browser, with an opt-in to your key. |
| 534 | assert.deepEqual(await ok("GET", `${P}/coordinator`), { enabled: false }); |
| 535 | const settings = { agent: "task-bot", runtime: "custom", command: "sh coord.sh", tokens_per_task: 150000, accept: true }; |
| 536 | assert.equal((await api("PUT", `${P}/coordinator`, settings)).status, 403); |
| 537 | assert.equal((await asPerson("PUT", `${P}/coordinator`, { ...settings, accept: false })).status, 422); |
| 538 | assert.equal((await asPerson("PUT", `${P}/coordinator`, { ...settings, tokens_per_task: 5 })).status, 422); |
| 539 | assert.equal((await asPerson("PUT", `${P}/coordinator`, { ...settings, agent: owner })).status, 422); |
| 540 | const saved = await asPerson("PUT", `${P}/coordinator`, settings); |
| 541 | assert.equal(saved.status, 200, JSON.stringify(saved.json)); |
| 542 | assert.equal(saved.json.agent.login, "task-bot"); |
| 543 | assert.equal(saved.json.tokens_per_task, 150000); |
| 544 | assert.equal(saved.json.tokens_per_run, 1000000); |
| 545 | // Everyone with write access sees whose coordinators are on, and with which agent and model; never their budgets or command. |
| 546 | const viewer = (await asPerson("POST", "/diffinery/agents/task-bot/tokens", { name: "coordinators" })).json.token; |
| 547 | const listed = await api("GET", `${P}/coordinators`, undefined, viewer); |
| 548 | assert.equal(listed.status, 200, JSON.stringify(listed.json)); |
| 549 | assert.equal(listed.json.length, 1, JSON.stringify(listed.json)); |
| 550 | assert.deepEqual(listed.json[0], { person: { login: owner }, agent: { login: "task-bot" }, runtime: "custom", provider: null, model: null, opted_in_at: listed.json[0].opted_in_at }); |
| 551 | // Without write access (the rival, signed in in a browser), the list doesn't exist. |
| 552 | const rivalList = await fetch(`${base}/api/v3${P}/coordinators`, { headers: { cookie: `__Host-dfy_session=${RIVAL_SESSION}` } }); |
| 553 | assert.equal(rivalList.status, 404, "people without write access see nothing"); |
| 554 | |
| 555 | const coordinatorRun = async (post) => |
| 556 | until(`the coordinator to handle post #${post}`, async () => { |
| 557 | const t = await ok("GET", `${R}/${post.task}`); |
| 558 | return t.runs.length >= post.run && t.runs[post.run - 1].state !== "boot" && !["queued", "boot", "install", "agent", "finish"].includes(t.runs[post.run - 1].state) ? t : null; |
| 559 | }); |
| 560 | |
| 561 | // A post goes to the poster's coordinator, which plans three tasks: the second after the first, the third after the second. |
| 562 | const plan = await asPerson("POST", `${P}/posts`, { body: "PLAN-THREE: split the work." }); |
| 563 | assert.equal(plan.status, 201, JSON.stringify(plan.json)); |
| 564 | assert.equal(plan.json.coordinator.status, "started"); |
| 565 | const coord = plan.json.coordinator.task; |
| 566 | const done = await coordinatorRun({ task: coord, run: 1 }); |
| 567 | const coordText = (await ok("GET", `${R}/${coord}/transcript`)).map((l) => l.text).join("\n"); |
| 568 | assert.equal(done.kind, "coordinator"); |
| 569 | assert.equal(done.runs[0].state, "done", coordText); |
| 570 | assert.match(coordText, new RegExp(`^--- post #${plan.json.id} by ${owner} ---$`, "m")); |
| 571 | assert.match(coordText, /^--- memory ---$/m, "the coordinator reads the memory"); |
| 572 | assert.match(coordText, /read-only sandbox/); |
| 573 | assert.match(coordText, /Applied: a reply, started #\d+, #\d+, #\d+, memory\./); |
| 574 | |
| 575 | const posts = await ok("GET", `${P}/posts?after=${plan.json.id - 1}`); |
| 576 | const reply = posts.find((p) => p.kind === "coordinator"); |
| 577 | assert.equal(reply.body, "Splitting it in three."); |
| 578 | assert.equal(reply.author.login, "task-bot"); |
| 579 | assert.equal(posts.find((p) => p.id === plan.json.id).handled_by.number, coord); |
| 580 | const started = posts.filter((p) => p.kind === "event" && p.body.startsWith(`${owner}'s coordinator started task`)); |
| 581 | assert.equal(started.length, 3, JSON.stringify(posts)); |
| 582 | const [a, b, c] = started.map((p) => p.task.number); |
| 583 | const memEvent = posts.find((p) => p.memory_version); |
| 584 | assert.match(memEvent.body, /changed the project memory to version \d+: 1 line added, 1 removed\./); |
| 585 | assert.equal((await ok("GET", `${P}/memory`)).body, "MEMORY-BY-COORD: parts are split in three."); |
| 586 | |
| 587 | const ta = await ok("GET", `${R}/${a}`); |
| 588 | assert.deepEqual(ta.started_via, { type: "coordinator", post: plan.json.id }); |
| 589 | assert.equal(ta.token_budget, 150000); |
| 590 | assert.equal(ta.title, "Small part"); |
| 591 | assert.equal(ta.goal, "COORD-SMALL: write the small part."); |
| 592 | assert.deepEqual((await ok("GET", `${R}/${b}`)).waits_for, [a]); |
| 593 | |
| 594 | // The first runs; the second starts only when the first is ready, and its budget stops it; the third never starts. |
| 595 | const ra = await until("the first part", async () => { |
| 596 | const t = await ok("GET", `${R}/${a}`); |
| 597 | return ["ready", "failed", "stopped"].includes(t.state) ? t : null; |
| 598 | }); |
| 599 | const textA = (await ok("GET", `${R}/${a}/transcript`)).map((l) => l.text).join("\n"); |
| 600 | assert.equal(ra.state, "ready", textA); |
| 601 | assert.match(textA, /model 200/); |
| 602 | assert.match(textA, new RegExp(`${owner}'s coordinator started the task`)); |
| 603 | const rb = await until("the second part", async () => { |
| 604 | const t = await ok("GET", `${R}/${b}`); |
| 605 | return ["ready", "failed", "stopped"].includes(t.state) ? t : null; |
| 606 | }); |
| 607 | const textB = (await ok("GET", `${R}/${b}/transcript`)).map((l) => l.text).join("\n"); |
| 608 | assert.equal(rb.state, "failed", textB); |
| 609 | assert.match(textB, /Queued: it starts when task #\d+ is ready\./); |
| 610 | assert.match(rb.error, /The run's next model request could take it past its budget of 150k tokens/); |
| 611 | assert.ok(Date.parse(rb.runs[0].started_at) >= Date.parse(ra.runs[0].ended_at)); |
| 612 | const rc = await until("the third part to be stopped", async () => { |
| 613 | const t = await ok("GET", `${R}/${c}`); |
| 614 | return t.state === "stopped" ? t : null; |
| 615 | }); |
| 616 | assert.equal(rc.error, `Task #${b}, which this task waits for, failed.`); |
| 617 | assert.deepEqual(rc.runs.map((r) => r.state), ["stopped"]); |
| 618 | |
| 619 | // An invalid batch is rejected whole: nothing is applied. |
| 620 | const memoryBefore = (await ok("GET", `${P}/memory`)).version; |
| 621 | const bad = await asPerson("POST", `${P}/posts`, { body: "BAD-BATCH: stop someone else's task." }); |
| 622 | assert.equal(bad.json.coordinator.task, coord); |
| 623 | const badRun = await coordinatorRun({ task: coord, run: 2 }); |
| 624 | assert.equal(badRun.runs[1].state, "failed"); |
| 625 | assert.match(badRun.runs[1].error, /Its actions were rejected: stop\[0\] must be one of the tasks you may stop\. Nothing was applied\./); |
| 626 | const afterBad = await ok("GET", `${P}/posts?after=${bad.json.id}`); |
| 627 | assert.ok(!afterBad.some((p) => p.kind === "coordinator"), "no reply was posted"); |
| 628 | assert.ok(afterBad.some((p) => p.body === `${owner}'s coordinator didn't finish handling post #${bad.json.id}: ${badRun.runs[1].error}`), JSON.stringify(afterBad)); |
| 629 | assert.equal((await ok("GET", `${P}/memory`)).version, memoryBefore); |
| 630 | // It may only message tasks it started (their runs have a budget), never the person's own hand-started ones. |
| 631 | const hand = await asPerson("POST", `${P}/posts`, { body: "MSG-HAND: nudge task one." }); |
| 632 | const handRun = await coordinatorRun({ task: coord, run: 3 }); |
| 633 | assert.match(handRun.runs[2].error, /messages\[0\]\.task must be one of the tasks you may message/); |
| 634 | assert.equal((await ok("GET", `${R}/1`)).runs.length, 2, "task #1 got no new run"); |
| 635 | void hand; |
| 636 | const own = await asPerson("POST", `${P}/posts`, { body: "MSG-OWN: follow up on your own task." }); |
| 637 | const ownRun = await coordinatorRun({ task: coord, run: 4 }); |
| 638 | assert.equal(ownRun.runs[3].state, "done", JSON.stringify(ownRun.runs[3])); |
| 639 | void own; |
| 640 | const followed = await until("the follow-up on the first part", async () => { |
| 641 | const t = await ok("GET", `${R}/${a}`); |
| 642 | return t.runs.length === 2 && !["queued", "boot", "install", "agent", "finish"].includes(t.runs[1].state) ? t : null; |
| 643 | }); |
| 644 | assert.equal(followed.runs[1].kind, "message"); |
| 645 | assert.match(followed.runs[1].message, new RegExp(`^\\(From ${owner}'s coordinator, handling post #\\d+\\.\\)\\nCheck it again\\.$`)); |
| 646 | |
| 647 | // A post can skip the coordinator, and be handed to it later, once. |
| 648 | const quiet = await asPerson("POST", `${P}/posts`, { body: "Just a note.", coordinator: false }); |
| 649 | assert.equal(quiet.json.coordinator, null); |
| 650 | assert.equal((await api("POST", `${P}/posts/${quiet.json.id}/handle`)).status, 403); |
| 651 | const handled = await asPerson("POST", `${P}/posts/${quiet.json.id}/handle`); |
| 652 | assert.equal(handled.status, 202, JSON.stringify(handled.json)); |
| 653 | assert.equal((await asPerson("POST", `${P}/posts/${quiet.json.id}/handle`)).status, 409); |
| 654 | const third = await coordinatorRun({ task: coord, run: 5 }); |
| 655 | assert.equal(third.runs[4].state, "done"); |
| 656 | assert.ok((await ok("GET", `${P}/posts?after=${quiet.json.id}`)).some((p) => p.kind === "coordinator" && p.body === "Noted, nothing to do.")); |
| 657 | // A coordinator takes no messages and isn't retried. |
| 658 | assert.equal((await asPerson("POST", `${R}/${coord}/messages`, { body: "hi" })).status, 409); |
| 659 | |
| 660 | // The board leaves the coordinator out; the pages show its work. |
| 661 | const board = await ok("GET", P); |
| 662 | assert.ok(!Object.values(board.board).flat().some((t) => t.number === coord)); |
| 663 | const page = await (await fetch(`${base}/${owner}/tasks-demo/project`)).text(); |
| 664 | assert.match(page, /Splitting it in three\./); |
| 665 | assert.match(page, /badge purple">Coordinator/); |
| 666 | assert.match(page, /The change to the memory/); |
| 667 | assert.match(page, /<tr class="add"><td class="ln">\+<\/td><td class="code">MEMORY-BY-COORD/); |
| 668 | assert.match(page, /Hand this to my coordinator/); |
| 669 | assert.doesNotMatch(page, /Other coordinators here/, "your own coordinator isn't listed again"); |
| 670 | assert.equal((await fetch(`${base}/${owner}/tasks-demo/project/coordinator`)).status, 200); |
| 671 | const taskPage = await (await fetch(`${base}/${owner}/tasks-demo/tasks/${a}`)).text(); |
| 672 | assert.match(taskPage, /coordinator started this task while handling/); |
| 673 | |
| 674 | // Turning it off: posts stay posts. |
| 675 | assert.equal((await asPerson("DELETE", `${P}/coordinator`)).status, 204); |
| 676 | assert.deepEqual(await ok("GET", `${P}/coordinator`), { enabled: false }); |
| 677 | const after = await asPerson("POST", `${P}/posts`, { body: "PLAN-THREE again" }); |
| 678 | assert.equal(after.json.coordinator, null); |
| 679 | assert.equal((await asPerson("POST", `${P}/posts/${after.json.id}/handle`)).status, 422); |
| 680 | }); |
| 681 | |
| 682 | test("plan first: a read-only plan run, then approve, change or split the plan", async () => { |
| 683 | const me = await ok("GET", "/user"); |
| 684 | const owner = me.login; |
| 685 | const P = `/repos/${owner}/tasks-demo/diffinery/project`; |
| 686 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 687 | const src = join(work, "src"); |
| 688 | writeFileSync( |
| 689 | join(src, "plan.sh"), |
| 690 | `p="$DIFFINERY_TASK_PROMPT" |
| 691 | case "$p" in |
| 692 | *"You are the project coordinator"*) |
| 693 | post=$(printf '%s\\n' "$p" | sed -n '/^--- post #/,/^--- end of post ---/p') |
| 694 | case "$post" in |
| 695 | *"Split the plan of task #"*"PLAN-V1-STEP"*) printf '%s\\n' '\`\`\`json' '{"reply": "Split it.", "tasks": [{"title": "Plan part one", "goal": "PLAN-PART-ONE: build step one."}]}' '\`\`\`' ;; |
| 696 | *) printf '%s\\n' '\`\`\`json' '{"reply": "Noted."}' '\`\`\`' ;; |
| 697 | esac ;; |
| 698 | *"You are planning task"*) |
| 699 | echo planned > should-not-land.txt |
| 700 | git add -A && git commit -qm "plan run commit" && git push -q origin HEAD:refs/heads/plan-run-push 2>/dev/null && echo "PUSHED" >&2 |
| 701 | case "$p" in |
| 702 | *"Your previous plan:"*"Use the KV store"*) printf '%s\\n' '## Plan' '1. PLAN-V2-STEP: write planned.txt, with the KV store.' ;; |
| 703 | *) printf '%s\\n' '## Plan' '1. PLAN-V1-STEP: write planned.txt.' ;; |
| 704 | esac ;; |
| 705 | *"approved. Follow it:"*) |
| 706 | printf '%s\\n' "$p" | sed -n '/^--- plan ---$/,/^--- end of plan ---$/p' |
| 707 | echo built > planned.txt ;; |
| 708 | *PLAN-PART-ONE*) echo part > part.txt ;; |
| 709 | *) echo other > other.txt ;; |
| 710 | esac |
| 711 | `, |
| 712 | ); |
| 713 | git(src, ["pull", "-q", "--rebase", "origin", "main"]); |
| 714 | git(src, ["add", "plan.sh"]); |
| 715 | git(src, ["commit", "-q", "-m", "plan script"]); |
| 716 | git(src, ["push", "-q", "origin", "main"]); |
| 717 | const settled = (n, what, fn = () => true) => |
| 718 | until(what, async () => { |
| 719 | const t = await ok("GET", `${R}/${n}`); |
| 720 | const last = t.runs.at(-1); |
| 721 | return last && !["queued", "boot", "install", "agent", "finish"].includes(last.state) && fn(t) ? t : null; |
| 722 | }); |
| 723 | const transcript = async (n) => (await ok("GET", `${R}/${n}/transcript`)).map((l) => l.text).join("\n"); |
| 724 | |
| 725 | // A plan-first task's first run plans in a read-only sandbox: nothing it changes is pushed. |
| 726 | assert.equal((await asPerson("POST", R, { goal: "x", runtime: "custom", command: "sh plan.sh", agent: "task-bot", plan_first: "yes" })).status, 422); |
| 727 | const started = await asPerson("POST", R, { goal: "PLANNED-GOAL: write planned.txt", runtime: "custom", command: "sh plan.sh", agent: "task-bot", plan_first: true }); |
| 728 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 729 | const n = started.json.number; |
| 730 | assert.equal(started.json.plan_first, true); |
| 731 | assert.equal(started.json.plan_state, "planning"); |
| 732 | const planned = await settled(n, "the plan"); |
| 733 | const text1 = await transcript(n); |
| 734 | assert.equal(planned.runs[0].kind, "plan"); |
| 735 | assert.equal(planned.runs[0].state, "done", text1); |
| 736 | assert.equal(planned.state, "ready"); |
| 737 | assert.equal(planned.plan_state, "waiting"); |
| 738 | assert.match(planned.plan, /^## Plan\n1\. PLAN-V1-STEP: write planned\.txt\.$/); |
| 739 | assert.equal(planned.pull_request, null); |
| 740 | assert.match(text1, /Starting a read-only sandbox to plan on main as task-bot\./); |
| 741 | assert.match(text1, /It plans first\./); |
| 742 | assert.ok(!text1.includes("PUSHED"), "the plan run's push was refused"); |
| 743 | assert.equal(git(src, ["ls-remote", "origin", "refs/heads/plan-run-push", `refs/heads/${planned.branch}`]).trim(), "", "nothing was pushed"); |
| 744 | const board = await ok("GET", P); |
| 745 | assert.ok(board.board.waiting.some((t) => t.number === n), "a plan waiting for approval waits on you"); |
| 746 | assert.ok((await ok("GET", `${P}/posts`)).some((p) => p.body === `Task #${n}'s plan needs approval.`)); |
| 747 | const page = await (await fetch(`${base}/${owner}/tasks-demo/tasks/${n}`)).text(); |
| 748 | assert.match(page, /PLAN-V1-STEP/); |
| 749 | assert.match(page, /Approve plan/); |
| 750 | |
| 751 | // Asking for changes plans again, with the plan it replaces. |
| 752 | assert.equal((await asPerson("POST", `${R}/${n}/messages`, { body: "Use the KV store" })).status, 201); |
| 753 | const replanned = await settled(n, "the second plan", (t) => t.runs.length === 2); |
| 754 | assert.equal(replanned.runs[1].kind, "plan"); |
| 755 | assert.equal(replanned.plan_state, "waiting"); |
| 756 | assert.match(replanned.plan, /PLAN-V2-STEP/); |
| 757 | assert.match(await transcript(n), /Planning again with the changes asked for\./); |
| 758 | |
| 759 | // Approving starts the build run, which carries the approved plan; only a person in a browser may approve. |
| 760 | assert.equal((await api("POST", `${R}/${n}/plan/approve`)).status, 403); |
| 761 | const approved = await asPerson("POST", `${R}/${n}/plan/approve`); |
| 762 | assert.equal(approved.status, 202, JSON.stringify(approved.json)); |
| 763 | assert.equal(approved.json.plan_state, "approved"); |
| 764 | assert.equal((await asPerson("POST", `${R}/${n}/plan/approve`)).status, 409); |
| 765 | const built = await settled(n, "the build", (t) => t.runs.length === 3); |
| 766 | const text3 = await transcript(n); |
| 767 | assert.equal(built.runs[2].kind, "build"); |
| 768 | assert.equal(built.runs[2].state, "done", text3); |
| 769 | assert.match(text3, /The plan is approved: building it\./); |
| 770 | assert.match(text3, /^--- plan ---\n## Plan\n1\. PLAN-V2-STEP: write planned\.txt, with the KV store\.\n--- end of plan ---$/m); |
| 771 | assert.ok(built.pull_request, "the build opened a pull request"); |
| 772 | assert.ok((await ok("GET", `${P}/posts`)).some((p) => p.body === `${owner} approved the plan of task #${n}.`)); |
| 773 | // A follow-up after approval is an ordinary run, still with the plan. |
| 774 | assert.equal((await asPerson("POST", `${R}/${n}/messages`, { body: "One more thing" })).status, 201); |
| 775 | const after = await settled(n, "the follow-up", (t) => t.runs.length === 4); |
| 776 | assert.equal(after.runs[3].kind, "message"); |
| 777 | |
| 778 | // A project can require plans: every task plans first, ticked or not, including ones coordinators start. |
| 779 | assert.equal((await asPerson("PUT", `${P}/settings`, { plan_first: "always" })).status, 422); |
| 780 | assert.equal((await asPerson("PUT", `${P}/settings`, { plan_first: "required" })).json.plan_first, "required"); |
| 781 | const forced = await asPerson("POST", R, { goal: "SPLIT-GOAL: a goal to split", runtime: "custom", command: "sh plan.sh", agent: "task-bot", plan_first: false }); |
| 782 | assert.equal(forced.status, 201, JSON.stringify(forced.json)); |
| 783 | const m = forced.json.number; |
| 784 | assert.equal(forced.json.plan_first, true); |
| 785 | await settled(m, "the plan to split", (t) => t.plan_state === "waiting"); |
| 786 | |
| 787 | // Splitting hands the plan to the person's coordinator; the tasks it starts link back, and the task is done with. |
| 788 | assert.equal((await asPerson("POST", `${R}/${m}/plan/split`)).status, 422, "no coordinator yet"); |
| 789 | const coord = await asPerson("PUT", `${P}/coordinator`, { agent: "task-bot", runtime: "custom", command: "sh plan.sh", accept: true }); |
| 790 | assert.equal(coord.status, 200, JSON.stringify(coord.json)); |
| 791 | assert.equal((await asPerson("POST", `${R}/${n}/plan/split`)).status, 409, "an approved plan can't be split"); |
| 792 | assert.equal((await api("POST", `${R}/${m}/plan/split`)).status, 403); |
| 793 | const split = await asPerson("POST", `${R}/${m}/plan/split`); |
| 794 | assert.equal(split.status, 202, JSON.stringify(split.json)); |
| 795 | assert.equal((await asPerson("POST", `${R}/${m}/plan/split`)).status, 409); |
| 796 | const splitPost = (await ok("GET", `${P}/posts?after=${split.json.post - 1}`)).find((p) => p.id === split.json.post); |
| 797 | assert.match(splitPost.body, new RegExp(`^Split the plan of task #${m} \\(SPLIT-GOAL: a goal to split\\) into tasks\\.`)); |
| 798 | assert.equal(splitPost.author.login, owner); |
| 799 | const child = await until("the coordinator to split the plan", async () => { |
| 800 | const posts = await ok("GET", `${P}/posts?after=${split.json.post}`); |
| 801 | const ev = posts.find((p) => p.kind === "event" && p.body.startsWith(`${owner}'s coordinator started task`)); |
| 802 | return ev ? ev.task.number : null; |
| 803 | }); |
| 804 | const c = await ok("GET", `${R}/${child}`); |
| 805 | assert.equal(c.parent_task, m); |
| 806 | assert.equal(c.title, "Plan part one"); |
| 807 | assert.equal(c.plan_first, true, "a coordinator's task plans first when the project requires it"); |
| 808 | const parent = await ok("GET", `${R}/${m}`); |
| 809 | assert.equal(parent.plan_state, "split"); |
| 810 | assert.ok((await ok("GET", P)).board.idle.some((t) => t.number === m)); |
| 811 | const childPage = await (await fetch(`${base}/${owner}/tasks-demo/tasks/${child}`)).text(); |
| 812 | assert.match(childPage, new RegExp(`Split from the plan of <a href="[^"]*/tasks/${m}">task #${m}</a>`)); |
| 813 | await settled(child, "the child's plan", (t) => t.plan_state === "waiting"); |
| 814 | |
| 815 | assert.equal((await asPerson("PUT", `${P}/settings`, { plan_first: "off" })).json.plan_first, "off"); |
| 816 | assert.equal((await asPerson("DELETE", `${P}/coordinator`)).status, 204); |
| 817 | }); |
| 818 | |
| 819 | test("best of N: attempts run side by side on their own branches, then one is picked", async () => { |
| 820 | const me = await ok("GET", "/user"); |
| 821 | const owner = me.login; |
| 822 | const P = `/repos/${owner}/tasks-demo/diffinery/project`; |
| 823 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 824 | const src = join(work, "src"); |
| 825 | // Attempts get the same prompt; each claims a slot (in the test's work directory, which the jails share) and does |
| 826 | // different work. |
| 827 | const slots = join(work, "best-slots"); |
| 828 | mkdirSync(slots, { recursive: true }); |
| 829 | writeFileSync( |
| 830 | join(src, "best.sh"), |
| 831 | `p="$DIFFINERY_TASK_PROMPT" |
| 832 | case "$p" in |
| 833 | *"New instructions from"*) echo follow > follow.txt; echo FOLLOWED-UP; exit 0 ;; |
| 834 | esac |
| 835 | tag=$(printf '%s\\n' "$p" | grep -o 'BEST-[a-z0-9]*' | head -1) |
| 836 | slot=0 |
| 837 | for i in 1 2 3 4; do if mkdir "${slots}/$tag-$i" 2>/dev/null; then slot=$i; break; fi; done |
| 838 | case $slot in |
| 839 | 1) echo one > one.txt; echo SUMMARY-ONE-FILE ;; |
| 840 | 2) echo a > a.txt; echo b > b.txt; echo c > c.txt; echo SUMMARY-THREE-FILES ;; |
| 841 | *) echo SLOT-FAIL; exit 3 ;; |
| 842 | esac |
| 843 | `, |
| 844 | ); |
| 845 | git(src, ["pull", "-q", "--rebase", "origin", "main"]); |
| 846 | git(src, ["add", "best.sh"]); |
| 847 | git(src, ["commit", "-q", "-m", "best of N script"]); |
| 848 | git(src, ["push", "-q", "origin", "main"]); |
| 849 | const tag = `BEST-${randomBytes(6).toString("hex")}`; |
| 850 | const start = { goal: `${tag}: try it three ways`, runtime: "custom", command: "sh best.sh", agent: "task-bot" }; |
| 851 | |
| 852 | for (const bad of [0, 5, 1.5, "x"]) assert.equal((await asPerson("POST", R, { ...start, attempts: bad })).status, 422, String(bad)); |
| 853 | const started = await asPerson("POST", R, { ...start, attempts: 3 }); |
| 854 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 855 | const n = started.json.number; |
| 856 | assert.equal(started.json.attempts, 3); |
| 857 | assert.equal(started.json.pick_state, "running"); |
| 858 | assert.deepEqual(started.json.runs.map((r) => r.attempt), [1, 2, 3]); |
| 859 | |
| 860 | // They run together, each on its own branch; the task waits on you when the last one ends. |
| 861 | const done = await until("the attempts", async () => { |
| 862 | const t = await ok("GET", `${R}/${n}`); |
| 863 | return !["running", "queued"].includes(t.state) ? t : null; |
| 864 | }); |
| 865 | const text = (await ok("GET", `${R}/${n}/transcript`)).map((l) => l.text).join("\n"); |
| 866 | assert.equal(done.state, "ready", text); |
| 867 | assert.equal(done.pick_state, "waiting"); |
| 868 | assert.equal(done.pull_request, null, "no pull request until a pick"); |
| 869 | assert.match(text, new RegExp(`Attempt 2 of 3, on ${done.branch}-a2\\.`)); |
| 870 | assert.ok((await ok("GET", P)).board.waiting.some((t) => t.number === n), "attempts to pick from wait on you"); |
| 871 | assert.ok((await ok("GET", `${P}/posts`)).some((p) => p.body === `Task #${n}'s 3 attempts are done: 2 of them pushed work to compare. Pick one.`)); |
| 872 | |
| 873 | const attempts = await ok("GET", `${R}/${n}/attempts`); |
| 874 | assert.equal(attempts.length, 3); |
| 875 | assert.deepEqual(attempts.map((a) => a.branch), [done.branch, `${done.branch}-a2`, `${done.branch}-a3`]); |
| 876 | const failed = attempts.find((a) => a.state === "failed"); |
| 877 | const one = attempts.find((a) => a.changes.files === 1); |
| 878 | const three = attempts.find((a) => a.changes.files === 3); |
| 879 | assert.ok(failed && one && three, JSON.stringify(attempts)); |
| 880 | assert.equal(failed.head_sha, null); |
| 881 | assert.match(one.summary, /SUMMARY-ONE-FILE/); |
| 882 | assert.match(three.summary, /SUMMARY-THREE-FILES/); |
| 883 | assert.deepEqual(three.changes, { files: 3, additions: 3, deletions: 0 }); |
| 884 | for (const a of [one, three]) assert.equal(git(src, ["ls-remote", "origin", `refs/heads/${a.branch}`]).trim().split(/\s/)[0], a.head_sha); |
| 885 | const page = await (await fetch(`${base}/${owner}/tasks-demo/tasks/${n}`)).text(); |
| 886 | assert.match(page, /<h2>Attempts<\/h2>/); |
| 887 | assert.match(page, new RegExp(`/tasks/${n}/attempts/${three.attempt}/pick`)); |
| 888 | |
| 889 | // Follow-ups wait for a pick; only the person who started it may pick, and only an attempt with work. |
| 890 | assert.equal((await asPerson("POST", `${R}/${n}/messages`, { body: "go on" })).status, 409); |
| 891 | assert.equal((await api("POST", `${R}/${n}/attempts/${three.attempt}/pick`)).status, 403); |
| 892 | assert.equal((await asPerson("POST", `${R}/${n}/attempts/${failed.attempt}/pick`)).status, 422); |
| 893 | assert.equal((await asPerson("POST", `${R}/${n}/attempts/9/pick`)).status, 404); |
| 894 | const picked = await asPerson("POST", `${R}/${n}/attempts/${three.attempt}/pick`); |
| 895 | assert.equal(picked.status, 202, JSON.stringify(picked.json)); |
| 896 | assert.equal(picked.json.picked_attempt, three.attempt); |
| 897 | assert.equal(picked.json.pick_state, "picked"); |
| 898 | assert.equal(picked.json.branch, three.branch); |
| 899 | assert.ok(picked.json.pull_request, "picking opens the pull request"); |
| 900 | const pr = await ok("GET", `/repos/${owner}/tasks-demo/pulls/${picked.json.pull_request.number}`); |
| 901 | assert.equal(pr.head.ref, three.branch); |
| 902 | assert.equal(git(src, ["ls-remote", "origin", `refs/heads/${one.branch}`]).trim(), "", "the other attempt's branch is deleted"); |
| 903 | assert.equal((await asPerson("POST", `${R}/${n}/attempts/${one.attempt}/pick`)).status, 409); |
| 904 | assert.ok((await ok("GET", `${P}/posts`)).some((p) => p.body === `${owner} picked attempt ${three.attempt} of task #${n}.`)); |
| 905 | |
| 906 | // Later runs build on the picked attempt. |
| 907 | assert.equal((await asPerson("POST", `${R}/${n}/messages`, { body: "Add a follow file" })).status, 201); |
| 908 | const next = await until("the follow-up", async () => { |
| 909 | const t = await ok("GET", `${R}/${n}`); |
| 910 | return t.runs.length === 4 && !["queued", "boot", "install", "agent", "finish"].includes(t.runs[3].state) ? t : null; |
| 911 | }); |
| 912 | assert.equal(next.runs[3].state, "done"); |
| 913 | assert.equal(next.runs[3].attempt, undefined); |
| 914 | assert.match((await ok("GET", `${R}/${n}/transcript`)).map((l) => l.text).join("\n"), new RegExp(`Starting a sandbox on ${three.branch} as task-bot\\.`)); |
| 915 | const files = (await ok("GET", `/repos/${owner}/tasks-demo/pulls/${pr.number}/files`)).map((f) => f.filename).sort(); |
| 916 | assert.deepEqual(files, ["a.txt", "b.txt", "c.txt", "follow.txt"]); |
| 917 | for (let i = 1; i <= 4; i++) rmSync(`/tmp/${tag}-${i}`, { recursive: true, force: true }); |
| 918 | }); |
| 919 | |
| 920 | test("verify and land: failing checks start a fix run, and an approved pull request lands", async () => { |
| 921 | const me = await ok("GET", "/user"); |
| 922 | const owner = me.login; |
| 923 | const P = `/repos/${owner}/tasks-demo/diffinery/project`; |
| 924 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 925 | const REPO = `/repos/${owner}/tasks-demo`; |
| 926 | const src = join(work, "src"); |
| 927 | writeFileSync( |
| 928 | join(src, "verify.sh"), |
| 929 | `p="$DIFFINERY_TASK_PROMPT" |
| 930 | case "$p" in |
| 931 | *"Checks failed on pull request"*) printf '%s\\n' "$p" | grep -E '^(--- |New instructions from|CI-SAYS)' ; echo fixed > fixed.txt ;; |
| 932 | *) echo work > verify.txt ;; |
| 933 | esac |
| 934 | `, |
| 935 | ); |
| 936 | git(src, ["pull", "-q", "--rebase", "origin", "main"]); |
| 937 | git(src, ["add", "verify.sh"]); |
| 938 | git(src, ["commit", "-q", "-m", "verify script"]); |
| 939 | git(src, ["push", "-q", "origin", "main"]); |
| 940 | // An unprotected branch to land on: agents may have pull requests merged into it automatically, once a person approves. |
| 941 | git(src, ["push", "-q", "origin", "main:refs/heads/release"]); |
| 942 | |
| 943 | const start = { goal: "VERIFY-GOAL: write verify.txt", runtime: "custom", command: "sh verify.sh", agent: "task-bot", base: "release" }; |
| 944 | assert.equal((await asPerson("POST", R, { ...start, land: "yes" })).status, 422); |
| 945 | const started = await asPerson("POST", R, { ...start, land: true }); |
| 946 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 947 | const n = started.json.number; |
| 948 | assert.equal(started.json.fix_checks, true, "on by default"); |
| 949 | assert.equal(started.json.land, true); |
| 950 | const ready = await until("the task's pull request", async () => { |
| 951 | const t = await ok("GET", `${R}/${n}`); |
| 952 | return t.state !== "running" && t.state !== "queued" ? t : null; |
| 953 | }); |
| 954 | assert.equal(ready.state, "ready", JSON.stringify(ready)); |
| 955 | const pr = ready.pull_request.number; |
| 956 | assert.match((await ok("GET", `${R}/${n}/transcript`)).map((l) => l.text).join("\n"), new RegExp(`Auto-merge is on for #${pr}: it lands once a person approves it and its checks pass\\.`)); |
| 957 | const pull = await ok("GET", `${REPO}/pulls/${pr}`); |
| 958 | assert.ok(pull.auto_merge, "auto-merge is on"); |
| 959 | assert.equal(pull.merged, false, "not without a person's approval"); |
| 960 | |
| 961 | // A failing check on the head starts one fix run, with what the check said. |
| 962 | const head = pull.head.sha; |
| 963 | await ok("POST", `${REPO}/statuses/${head}`, { state: "failure", context: "ci/verify", description: "CI-SAYS: verify.txt is wrong" }); |
| 964 | const fixing = await until("the fix run", async () => { |
| 965 | const t = await ok("GET", `${R}/${n}`); |
| 966 | return t.runs.length === 2 && !["queued", "boot", "install", "agent", "finish"].includes(t.runs[1].state) ? t : null; |
| 967 | }); |
| 968 | const text = (await ok("GET", `${R}/${n}/transcript`)).map((l) => l.text).join("\n"); |
| 969 | assert.equal(fixing.runs[1].kind, "fix"); |
| 970 | assert.equal(fixing.runs[1].state, "done", text); |
| 971 | assert.equal(fixing.check_fixes, 1); |
| 972 | assert.match(text, /^--- ci\/verify: failure ---$/m); |
| 973 | assert.match(text, /^CI-SAYS: verify\.txt is wrong$/m); |
| 974 | assert.match(text, /^New instructions from Diffinery \(the pull request's checks failed\):$/m); |
| 975 | assert.ok((await ok("GET", `${P}/posts`)).some((p) => p.body === `Task #${n}'s pull request #${pr} failed checks; starting fix run 1 of 3.`)); |
| 976 | const fixedHead = (await ok("GET", `${REPO}/pulls/${pr}`)).head.sha; |
| 977 | assert.notEqual(fixedHead, head); |
| 978 | // The same failure reported again on the old head starts nothing more. |
| 979 | await ok("POST", `${REPO}/statuses/${head}`, { state: "failure", context: "ci/other" }); |
| 980 | await ok("POST", `${REPO}/statuses/${fixedHead}`, { state: "success", context: "ci/verify" }); |
| 981 | await sleep(3000); |
| 982 | assert.equal((await ok("GET", `${R}/${n}`)).runs.length, 2); |
| 983 | |
| 984 | // Options: only the person who started it, in a browser. |
| 985 | assert.equal((await api("PATCH", `${R}/${n}`, { land: false })).status, 403); |
| 986 | assert.equal((await asPerson("PATCH", `${R}/${n}`, { fix_checks: "no" })).status, 422); |
| 987 | const off = await asPerson("PATCH", `${R}/${n}`, { fix_checks: false }); |
| 988 | assert.equal(off.status, 200, JSON.stringify(off.json)); |
| 989 | assert.equal(off.json.fix_checks, false); |
| 990 | assert.match(await (await fetch(`${base}/${owner}/tasks-demo/tasks/${n}`)).text(), /<h2>Verify and land<\/h2>/); |
| 991 | |
| 992 | // A person approves: it lands by itself. |
| 993 | const review = await api("POST", `${REPO}/pulls/${pr}/reviews`, { event: "APPROVE", commit_id: fixedHead }); |
| 994 | assert.ok(review.status < 300, JSON.stringify(review.json)); |
| 995 | const merged = await until("the pull request to land", async () => { |
| 996 | const p = await ok("GET", `${REPO}/pulls/${pr}`); |
| 997 | return p.merged ? p : null; |
| 998 | }); |
| 999 | assert.equal(merged.merged, true); |
| 1000 | assert.ok((await ok("GET", P)).board.done.some((t) => t.number === n), "a landed task is done"); |
| 1001 | }); |
| 1002 | |
| 1003 | test("organizations: an agent policy, model keys only owners hold, and whose key pays (phase 5b)", async () => { |
| 1004 | const me = await ok("GET", "/user"); |
| 1005 | const owner = me.login; |
| 1006 | const ORG = "acme-agents"; |
| 1007 | const O = `/orgs/${ORG}/diffinery`; |
| 1008 | const ORG_KEY = `sk-ant-api03-${randomBytes(30).toString("hex")}`; |
| 1009 | const RIVAL_KEY = `sk-ant-api03-${randomBytes(30).toString("hex")}`; |
| 1010 | const rivalCookie = `__Host-dfy_session=${RIVAL_SESSION}`; |
| 1011 | const asRival = async (method, path, body) => { |
| 1012 | const res = await fetch(`${base}/api/v3${path}`, { method, headers: { cookie: rivalCookie, "content-type": "application/json" }, body: body === undefined ? undefined : JSON.stringify(body) }); |
| 1013 | const text = await res.text(); |
| 1014 | let json = null; |
| 1015 | try { |
| 1016 | json = text ? JSON.parse(text) : null; |
| 1017 | } catch { |
| 1018 | json = text; |
| 1019 | } |
| 1020 | return { status: res.status, json }; |
| 1021 | }; |
| 1022 | // A run's request through the model proxy, with its token, as the agent's CLI would send it. |
| 1023 | const proxyBody = (tok, body) => fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", headers: { "x-api-key": tok, "content-type": "application/json" }, body }); |
| 1024 | const proxy = (tok, max = 100) => proxyBody(tok, JSON.stringify({ model: "m", max_tokens: max, messages: [] })); |
| 1025 | // A web form, as the owner in their browser. |
| 1026 | const form = async (page, action, fields) => { |
| 1027 | const res = await fetch(`${base}${page}`); |
| 1028 | const csrf = /name="_csrf" value="([^"]+)"/.exec(await res.text())?.[1]; |
| 1029 | const cookie = /dfy_csrf=([^;]+)/.exec(res.headers.get("set-cookie") ?? "")?.[1]; |
| 1030 | return fetch(`${base}${action}`, { method: "POST", redirect: "manual", headers: { cookie: `dfy_csrf=${cookie}`, "sec-fetch-site": "same-origin" }, body: new URLSearchParams([["_csrf", csrf ?? ""], ...fields]) }); |
| 1031 | }; |
| 1032 | const runToken = (file) => jailToken(join(work, file)); |
| 1033 | const stopped = (R, n) => |
| 1034 | until(`task #${n} to stop`, async () => { |
| 1035 | const t = await ok("GET", `${R}/${n}`); |
| 1036 | return t.state !== "running" && t.state !== "queued" ? t : null; |
| 1037 | }); |
| 1038 | |
| 1039 | await ok("POST", "/user/orgs", { login: ORG }); |
| 1040 | await ok("POST", `/orgs/${ORG}/repos`, { name: "org-tasks" }); |
| 1041 | const src = join(work, "org-src"); |
| 1042 | mkdirSync(src, { recursive: true }); |
| 1043 | writeFileSync(join(src, "README.md"), "# org tasks\n"); |
| 1044 | git(src, ["init", "-q", "-b", "main"]); |
| 1045 | git(src, ["add", "-A"]); |
| 1046 | git(src, ["commit", "-q", "-m", "initial"]); |
| 1047 | git(src, ["remote", "add", "origin", `${base}/${ORG}/org-tasks.git`]); |
| 1048 | git(src, ["push", "-q", "origin", "main"]); |
| 1049 | assert.equal((await asPerson("PUT", "/diffinery/model-keys/anthropic", { key: ANTHROPIC_KEY, accept_terms: true })).status, 200); |
| 1050 | assert.equal((await asPerson("POST", "/diffinery/agents", { login: "org-bot" })).status, 201); |
| 1051 | |
| 1052 | // Nothing changes until an owner sets something; only owners see it, and never with a repository-limited token. |
| 1053 | assert.deepEqual(await ok("GET", `${O}/agent-policy`), { keys: "personal", runtimes: ["claude-code", "codex", "custom"], providers: ["anthropic", "openai", "openrouter"], max_running: 10, terminals: true, terminal_tasks: true, review: true, member_monthly_tokens: null, updated_by: null, updated_at: null }); |
| 1054 | assert.equal((await asRival("GET", `${O}/agent-policy`)).status, 404, "not a member"); |
| 1055 | const limited = (await asPerson("POST", "/diffinery/tokens", { name: "org-limited", scopes: ["admin"], repositories: [`${ORG}/org-tasks`] })).json.token; |
| 1056 | for (const path of ["agent-policy", "model-keys", "model-usage"]) assert.equal((await api("GET", `${O}/${path}`, undefined, limited)).status, 403, path); |
| 1057 | |
| 1058 | // The organization's key: an owner in a browser, accepting the terms for it; never shown again. |
| 1059 | const viaToken = await api("PUT", `${O}/model-keys/anthropic`, { key: ORG_KEY, accept_terms: true }); |
| 1060 | assert.equal(viaToken.status, 403, JSON.stringify(viaToken.json)); |
| 1061 | assert.equal((await asPerson("PUT", `${O}/model-keys/anthropic`, { key: ORG_KEY })).status, 422); |
| 1062 | const saved = await asPerson("PUT", `${O}/model-keys/anthropic`, { key: ORG_KEY, accept_terms: true }); |
| 1063 | assert.equal(saved.status, 200, JSON.stringify(saved.json)); |
| 1064 | assert.equal(saved.json.hint, ORG_KEY.slice(-4)); |
| 1065 | assert.equal(saved.json.opted_in_by, owner); |
| 1066 | const keys = await ok("GET", `${O}/model-keys`); |
| 1067 | assert.deepEqual(keys.map((k) => k.provider), ["anthropic"]); |
| 1068 | assert.ok(!JSON.stringify(keys).includes(ORG_KEY)); |
| 1069 | assert.ok(!(await ok("GET", "/diffinery/model-keys")).some((k) => k.hint === ORG_KEY.slice(-4)), "a person's keys are their own"); |
| 1070 | |
| 1071 | // The policy: owners in a browser. |
| 1072 | assert.equal((await api("PUT", `${O}/agent-policy`, { keys: "organization" })).status, 403, "a token can't change it"); |
| 1073 | assert.equal((await asPerson("PUT", `${O}/agent-policy`, { max_running: 11 })).status, 422); |
| 1074 | const policy = await asPerson("PUT", `${O}/agent-policy`, { keys: "organization", runtimes: ["claude-code", "custom"], max_running: 2, member_monthly_tokens: 10000 }); |
| 1075 | assert.equal(policy.status, 200, JSON.stringify(policy.json)); |
| 1076 | assert.equal(policy.json.updated_by, owner); |
| 1077 | // Its repositories' "tasks at once" is held to it. |
| 1078 | await until("the organization's limit to reach the project", async () => (await ok("GET", `/repos/${ORG}/org-tasks/diffinery/project`)).settings.org_max_running === 2); |
| 1079 | |
| 1080 | const R = `/repos/${ORG}/org-tasks/diffinery/tasks`; |
| 1081 | const codex = await asPerson("POST", R, { goal: "Not allowed", runtime: "codex", agent: "org-bot" }); |
| 1082 | assert.equal(codex.status, 403, JSON.stringify(codex.json)); |
| 1083 | assert.match(codex.json.message, /acme-agents's agent policy doesn't allow Codex/); |
| 1084 | |
| 1085 | // An owner's task runs on the organization's key, which the provider gets and the sandbox never sees. |
| 1086 | const started = await asPerson("POST", R, { goal: "On the organization's key", runtime: "custom", command: `printf %s "$ANTHROPIC_API_KEY" > ${join(work, "org-token")}; echo "key is $ANTHROPIC_API_KEY"; sleep 120`, agent: "org-bot" }); |
| 1087 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 1088 | assert.equal(started.json.paid_by, "organization"); |
| 1089 | const n1 = started.json.number; |
| 1090 | const t1 = await runToken("org-token"); |
| 1091 | assert.match(t1, /^dmt_/); |
| 1092 | const sent = modelRequests.length; |
| 1093 | const first = await proxy(t1); |
| 1094 | assert.equal(first.status, 200, await first.clone().text()); |
| 1095 | assert.equal(modelRequests.length, sent + 1); |
| 1096 | assert.equal(modelRequests.at(-1).headers["x-api-key"], ORG_KEY); |
| 1097 | // The month's count settles to what the answer used: 21 + 4 + 300 cache reads at a tenth. |
| 1098 | const month = await until("the month's count", async () => { |
| 1099 | const u = await ok("GET", `${O}/model-usage`); |
| 1100 | return u.month.people.find((p) => p.person === owner && p.tokens === 55) ? u : null; |
| 1101 | }); |
| 1102 | assert.equal(month.month.member_monthly_tokens, 10000); |
| 1103 | // A request that could take the person past the allowance never reaches the provider, and the task stops. |
| 1104 | const over = await proxy(t1, 20000); |
| 1105 | assert.equal(over.status, 403); |
| 1106 | assert.match((await over.json()).error.message, /allowance of 10,000 tokens per person on acme-agents's keys is used up/); |
| 1107 | assert.equal(modelRequests.length, sent + 1); |
| 1108 | assert.equal((await stopped(R, n1)).state, "stopped"); |
| 1109 | const lines = (await ok("GET", `${R}/${n1}/transcript`)).map((l) => l.text).join("\n"); |
| 1110 | assert.match(lines, /Stopped: this month's allowance/); |
| 1111 | assert.ok(!lines.includes(ORG_KEY)); |
| 1112 | // The run is the organization's to see, not in the owner's own 30 days. |
| 1113 | const orgReport = await until("the organization's report", async () => { |
| 1114 | const u = await ok("GET", `${O}/model-usage`); |
| 1115 | return u.repositories.find((x) => x.repository === `${ORG}/org-tasks`) ? u : null; |
| 1116 | }); |
| 1117 | assert.deepEqual(orgReport.repositories[0], { repository: `${ORG}/org-tasks`, kind: "task", runs: 1, reported_by_cli_runs: 0, requests: 1, input_tokens: 21, output_tokens: 4, cache_read_tokens: 300, cache_write_tokens: 0, cost_usd: null, person: owner }); |
| 1118 | assert.ok(!(await ok("GET", "/diffinery/model-usage")).repositories.some((x) => x.repository === `${ORG}/org-tasks`)); |
| 1119 | |
| 1120 | // Someone outside the organization, with write access, their own agent and their own key. |
| 1121 | await ok("PUT", `/repos/${ORG}/org-tasks/collaborators/rival`, { permission: "push" }); |
| 1122 | const inv = (await asRival("GET", "/user/repository_invitations")).json.find((i) => i.repository.full_name === `${ORG}/org-tasks`); |
| 1123 | assert.equal((await asRival("PATCH", `/user/repository_invitations/${inv.id}`)).status, 204); |
| 1124 | assert.equal((await asRival("PUT", "/diffinery/model-keys/anthropic", { key: RIVAL_KEY, accept_terms: true })).status, 200); |
| 1125 | assert.equal((await asRival("POST", "/diffinery/agents", { login: "rival-bot" })).status, 201); |
| 1126 | const outside = await asRival("POST", R, { goal: "Outside", runtime: "custom", command: "true", agent: "rival-bot" }); |
| 1127 | assert.equal(outside.status, 403, JSON.stringify(outside.json)); |
| 1128 | assert.match(outside.json.message, /only its members can use/); |
| 1129 | // Under `either`, they pay with their own key. |
| 1130 | assert.equal((await asPerson("PUT", `${O}/agent-policy`, { keys: "either", member_monthly_tokens: null })).status, 200); |
| 1131 | const own = await asRival("POST", R, { goal: "Own key", runtime: "custom", command: `printf %s "$ANTHROPIC_API_KEY" > ${join(work, "rival-token")}; sleep 120`, agent: "rival-bot" }); |
| 1132 | assert.equal(own.status, 201, JSON.stringify(own.json)); |
| 1133 | assert.equal(own.json.paid_by, "person"); |
| 1134 | const t2 = await runToken("rival-token"); |
| 1135 | assert.equal((await proxy(t2)).status, 200); |
| 1136 | assert.equal(modelRequests.at(-1).headers["x-api-key"], RIVAL_KEY); |
| 1137 | // On their own key, a request goes as they sent it. |
| 1138 | assert.equal((await proxyBody(t2, '{"model":"m","background":true,"messages":[]}')).status, 200); |
| 1139 | assert.deepEqual(JSON.parse(modelRequests.at(-1).body), { model: "m", background: true, messages: [] }); |
| 1140 | assert.equal((await asRival("POST", `${R}/${own.json.number}/stop`)).status, 200); |
| 1141 | |
| 1142 | // Once a member, their task runs on the organization's key, until they're removed: the next request is refused. |
| 1143 | const invited = await form(`/${ORG}?tab=people`, `/${ORG}/-/org/people`, [["who", "rival"], ["role", "member"]]); |
| 1144 | assert.equal(invited.status, 303); |
| 1145 | assert.equal((await asRival("PATCH", `/user/memberships/orgs/${ORG}`, { state: "active" })).status, 200); |
| 1146 | const member = await asRival("POST", R, { goal: "As a member", runtime: "custom", command: `printf %s "$ANTHROPIC_API_KEY" > ${join(work, "member-token")}; sleep 120`, agent: "rival-bot" }); |
| 1147 | assert.equal(member.status, 201, JSON.stringify(member.json)); |
| 1148 | assert.equal(member.json.paid_by, "organization"); |
| 1149 | const t3 = await runToken("member-token"); |
| 1150 | assert.equal((await proxy(t3)).status, 200); |
| 1151 | assert.equal(modelRequests.at(-1).headers["x-api-key"], ORG_KEY); |
| 1152 | // On the organization's key, a run sends only what its booking bounds (Security M1): no background responses, |
| 1153 | // stored context or several answers, and its output capped as booked. |
| 1154 | const sentBefore = modelRequests.length; |
| 1155 | const background = await proxyBody(t3, '{"model":"m","max_tokens":100,"background":true,"messages":[]}'); |
| 1156 | assert.equal(background.status, 403); |
| 1157 | assert.match((await background.json()).error.message, /background responses aren't available to agent reviews, runs with a token budget and runs on an organization's key/); |
| 1158 | assert.equal((await proxyBody(t3, '{"model":"m","max_tokens":100,"n":3,"messages":[]}')).status, 403); |
| 1159 | // Nor input the request doesn't carry (Security's re-check): a PDF by URL, a stored item, a body that isn't JSON. |
| 1160 | const byUrl = await proxyBody(t3, JSON.stringify({ model: "m", max_tokens: 16, messages: [{ role: "user", content: [{ type: "document", source: { type: "url", url: "https://example.com/big.pdf" } }] }] })); |
| 1161 | assert.equal(byUrl.status, 403); |
| 1162 | assert.match((await byUrl.json()).error.message, /^inputs given by URL aren't available to agent reviews, runs with a token budget and runs on an organization's key/); |
| 1163 | assert.equal((await proxyBody(t3, '{"model":"m","max_tokens":16,"background":"true","messages":[]}')).status, 403); |
| 1164 | assert.equal((await proxyBody(t3, '{"model":"m","max_tokens":16,"temperature":NaN,"messages":[]}')).status, 403); |
| 1165 | assert.equal(modelRequests.length, sentBefore); |
| 1166 | // Its tools that run on the provider are stripped; those the agent runs stay. |
| 1167 | assert.equal((await proxyBody(t3, JSON.stringify({ model: "m", messages: [], tools: [{ name: "Read", input_schema: {} }, { type: "web_fetch_20250910", name: "web_fetch" }] }))).status, 200); |
| 1168 | assert.deepEqual(JSON.parse(modelRequests.at(-1).body), { model: "m", messages: [], tools: [{ name: "Read", input_schema: {} }], max_tokens: 32000 }); |
| 1169 | // An absurd cap books a bounded estimate (Security S2): the month's count stays exact, here 3 answers of 55. |
| 1170 | assert.equal((await proxyBody(t3, '{"model":"m","max_tokens":1e300,"messages":[]}')).status, 200); |
| 1171 | assert.equal(JSON.parse(modelRequests.at(-1).body).max_tokens, 1000000); |
| 1172 | await until("the member's exact count", async () => (await ok("GET", `${O}/model-usage`)).month.people.find((p) => p.person === "rival" && p.tokens === 165)); |
| 1173 | // Under an allowance, a person has at most 4 requests on the organization's keys on their way at once (Security's third |
| 1174 | // look: images and PDFs bill past their booking, so this bounds the overshoot); a fifth is told to try again. |
| 1175 | assert.equal((await asPerson("PUT", `${O}/agent-policy`, { member_monthly_tokens: 1_000_000 })).status, 200); |
| 1176 | const slow = await Promise.all([1, 2, 3, 4, 5].map((i) => proxyBody(t3, JSON.stringify({ model: "m", max_tokens: 10, metadata: { user_id: "slow" }, messages: [{ role: "user", content: `slow ${i}` }] })))); |
| 1177 | assert.deepEqual(slow.map((r) => r.status).sort(), [200, 200, 200, 200, 429]); |
| 1178 | const fifth = slow.find((r) => r.status === 429); |
| 1179 | assert.equal(fifth.headers.get("retry-after"), "5"); |
| 1180 | assert.match((await fifth.json()).error.message, /4 of your model requests on acme-agents's keys are already on their way/); |
| 1181 | // Once they are answered, the slots are free again. |
| 1182 | await until("the slots to free up", async () => (await proxy(t3)).status === 200); |
| 1183 | // A request whose sandbox hangs up while the provider is still answering is settled anyway, and frees its slot |
| 1184 | // (Security's fourth look): the next 4 at once all go through. |
| 1185 | const hangUp = new AbortController(); |
| 1186 | const dropped = fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", signal: hangUp.signal, headers: { "x-api-key": t3, "content-type": "application/json" }, body: JSON.stringify({ model: "m", max_tokens: 10, metadata: { user_id: "slow" }, messages: [] }) }).catch(() => null); |
| 1187 | await sleep(1000); |
| 1188 | hangUp.abort(); |
| 1189 | await dropped; |
| 1190 | await sleep(4000); |
| 1191 | const after4 = await Promise.all([1, 2, 3, 4].map(() => proxyBody(t3, JSON.stringify({ model: "m", max_tokens: 10, metadata: { user_id: "slow" }, messages: [] })))); |
| 1192 | assert.deepEqual(after4.map((r) => r.status), [200, 200, 200, 200]); |
| 1193 | assert.equal((await asPerson("PUT", `${O}/agent-policy`, { member_monthly_tokens: null })).status, 200); |
| 1194 | const removed = await form(`/${ORG}?tab=people`, `/${ORG}/-/org/members`, [["login", "rival"], ["role", "remove"]]); |
| 1195 | assert.equal(removed.status, 303); |
| 1196 | // Removal also takes away their grants on its repositories; with write access given back as an outside |
| 1197 | // collaborator, they can still run the task, but not on the organization's key. |
| 1198 | await ok("PUT", `/repos/${ORG}/org-tasks/collaborators/rival`, { permission: "push" }); |
| 1199 | const again = (await asRival("GET", "/user/repository_invitations")).json.find((i) => i.repository.full_name === `${ORG}/org-tasks`); |
| 1200 | if (again) assert.equal((await asRival("PATCH", `/user/repository_invitations/${again.id}`)).status, 204); |
| 1201 | const before = modelRequests.length; |
| 1202 | const refused = await proxy(t3); |
| 1203 | assert.equal(refused.status, 403); |
| 1204 | assert.match((await refused.json()).error.message, /rival is no longer a member of acme-agents, whose key this task uses/); |
| 1205 | assert.equal(modelRequests.length, before); |
| 1206 | assert.equal((await stopped(R, member.json.number)).state, "stopped"); |
| 1207 | // Retrying it isn't theirs to pay with the organization's key any more. |
| 1208 | const retry = await asRival("POST", `${R}/${member.json.number}/retry`); |
| 1209 | assert.equal(retry.status, 403, JSON.stringify(retry.json)); |
| 1210 | assert.match(retry.json.message, /only members of acme-agents can run agents on its keys/); |
| 1211 | |
| 1212 | // The Settings tab: owners see the policy, the key's last four characters and the usage; nobody sees the key. |
| 1213 | const page = await (await fetch(`${base}/${ORG}?tab=settings`)).text(); |
| 1214 | assert.match(page, /Agent policy/); |
| 1215 | assert.ok(page.includes(`…${ORG_KEY.slice(-4)}`)); |
| 1216 | assert.ok(!page.includes(ORG_KEY)); |
| 1217 | assert.match(page, /Paid with the organization/); |
| 1218 | assert.equal((await fetch(`${base}/${ORG}?tab=settings`, { headers: { cookie: rivalCookie } })).status, 403); |
| 1219 | // Its form saves the policy (Kumo checkboxes send only what's ticked). |
| 1220 | const savedForm = await form(`/${ORG}?tab=settings`, `/${ORG}/-/org/agent-policy`, [ |
| 1221 | ["keys", "personal"], |
| 1222 | ["runtimes", "claude-code"], |
| 1223 | ["runtimes", "custom"], |
| 1224 | ["providers", "anthropic"], |
| 1225 | ["max_running", "3"], |
| 1226 | ["terminal_tasks", "1"], |
| 1227 | ["review", "1"], |
| 1228 | ["member_monthly_tokens", ""], |
| 1229 | ]); |
| 1230 | assert.equal(savedForm.status, 303); |
| 1231 | assert.deepEqual(await ok("GET", `${O}/agent-policy`), { ...(await ok("GET", `${O}/agent-policy`)), keys: "personal", runtimes: ["claude-code", "custom"], providers: ["anthropic"], max_running: 3, terminals: false, terminal_tasks: true, review: true, member_monthly_tokens: null }); |
| 1232 | // Deleting the key. |
| 1233 | assert.equal((await asPerson("DELETE", `${O}/model-keys/anthropic`)).status, 204); |
| 1234 | assert.deepEqual(await ok("GET", `${O}/model-keys`), []); |
| 1235 | }); |
| 1236 | |
| 1237 | test("losing access to the agent stops the task", async () => { |
| 1238 | const me = await ok("GET", "/user"); |
| 1239 | const owner = me.login; |
| 1240 | const R = `/repos/${owner}/tasks-demo/diffinery/tasks`; |
| 1241 | const tokFile = join(work, "model-token-3"); |
| 1242 | const started = await asPerson("POST", R, { goal: "Lose access", runtime: "custom", command: `printf %s "$ANTHROPIC_API_KEY" > ${tokFile}; sleep 120`, agent: "task-bot" }); |
| 1243 | assert.equal(started.status, 201, JSON.stringify(started.json)); |
| 1244 | const n = started.json.number; |
| 1245 | const runToken = await jailToken(tokFile); |
| 1246 | // Disabling the agent ends its access, and with it the task's. |
| 1247 | assert.equal((await api("DELETE", "/diffinery/agents/task-bot")).status, 204); |
| 1248 | const refused = await fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", headers: { "x-api-key": runToken, "content-type": "application/json" }, body: "{}" }); |
| 1249 | assert.equal(refused.status, 403); |
| 1250 | const t = await until("the task to stop", async () => { |
| 1251 | const x = await ok("GET", `${R}/${n}`); |
| 1252 | return x.state !== "running" ? x : null; |
| 1253 | }); |
| 1254 | assert.equal(t.state, "stopped"); |
| 1255 | assert.equal((await fetch(`${base}/-/model/anthropic/v1/messages`, { method: "POST", headers: { "x-api-key": runToken }, body: "{}" })).status, 401); |
| 1256 | }); |