CodeIssuesDiscussionsWikiPull RequestsProjectsCommitsActionsReleasesContributorsPulse● GatesSecuritySettingsDeploymentsPipelineInsightsAgents✨ Explain✨ Ask AI✨ Workspace✨ Spec✨ Tests▓ Debt Map✨ NL Search🏛 Archaeology
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 | /**
* Block K1 — Agent runtime unit tests.
*
* Covers the pure helpers (state machine + log truncation) plus the
* runSandboxed primitive via real subprocesses. DB helpers are intentionally
* not tested here — they hit Neon and are covered by integration flow.
*/
import { describe, it, expect } from "bun:test";
import {
type AgentRunStatus,
canTransition,
isTerminalStatus,
runSandboxed,
truncateError,
truncateLog,
__internal,
} from "../lib/agent-runtime";
const ALL_STATUSES: AgentRunStatus[] = [
"queued",
"running",
"succeeded",
"failed",
"killed",
"timeout",
];
describe("agent-runtime — isTerminalStatus", () => {
it("treats queued + running as non-terminal", () => {
expect(isTerminalStatus("queued")).toBe(false);
expect(isTerminalStatus("running")).toBe(false);
});
it("treats succeeded / failed / killed / timeout as terminal", () => {
expect(isTerminalStatus("succeeded")).toBe(true);
expect(isTerminalStatus("failed")).toBe(true);
expect(isTerminalStatus("killed")).toBe(true);
expect(isTerminalStatus("timeout")).toBe(true);
});
it("is exhaustive over AgentRunStatus", () => {
// Every status value must have a defined answer (no undefined bubbling).
for (const s of ALL_STATUSES) {
expect(typeof isTerminalStatus(s)).toBe("boolean");
}
});
});
describe("agent-runtime — canTransition", () => {
it("allows the happy path queued → running → succeeded", () => {
expect(canTransition("queued", "running")).toBe(true);
expect(canTransition("running", "succeeded")).toBe(true);
});
it("allows running → failed | timeout | killed", () => {
expect(canTransition("running", "failed")).toBe(true);
expect(canTransition("running", "timeout")).toBe(true);
expect(canTransition("running", "killed")).toBe(true);
});
it("allows queued → killed (operator cancel before start)", () => {
expect(canTransition("queued", "killed")).toBe(true);
});
it("forbids going backward from any terminal state", () => {
for (const terminal of ["succeeded", "failed", "killed", "timeout"] as const) {
for (const to of ALL_STATUSES) {
expect(canTransition(terminal, to)).toBe(false);
}
}
});
it("forbids self-transitions", () => {
for (const s of ALL_STATUSES) {
expect(canTransition(s, s)).toBe(false);
}
});
it("forbids skipping running (queued → succeeded/failed/timeout)", () => {
expect(canTransition("queued", "succeeded")).toBe(false);
expect(canTransition("queued", "failed")).toBe(false);
expect(canTransition("queued", "timeout")).toBe(false);
});
it("forbids running → queued (no re-queue)", () => {
expect(canTransition("running", "queued")).toBe(false);
});
});
describe("agent-runtime — truncateLog", () => {
it("returns a concatenation when well under the cap", () => {
const out = truncateLog("abc", "def", 1024);
expect(out).toBe("abcdef");
});
it("returns exactly the combined string at the cap boundary", () => {
const existing = "a".repeat(500);
const addition = "b".repeat(500);
const out = truncateLog(existing, addition, 1000);
expect(out.length).toBe(1000);
expect(out).toBe(existing + addition);
});
it("prepends the sentinel and keeps the last maxBytes when over the cap", () => {
const existing = "x".repeat(2000);
const addition = "y".repeat(500);
const out = truncateLog(existing, addition, 1000);
expect(out.startsWith(__internal.LOG_TRUNCATED_SENTINEL)).toBe(true);
// Tail length == maxBytes; we keep the rightmost characters of existing+addition.
const tail = out.slice(__internal.LOG_TRUNCATED_SENTINEL.length);
expect(tail.length).toBe(1000);
// Since addition ('y'*500) sits at the very end, the last 500 chars must be 'y'.
expect(tail.endsWith("y".repeat(500))).toBe(true);
// And the middle should be the tail of the 'x' run.
expect(tail.startsWith("x".repeat(500))).toBe(true);
});
it("does not double the sentinel on subsequent truncations", () => {
// First truncation.
const first = truncateLog("x".repeat(2000), "", 1000);
expect(first.startsWith(__internal.LOG_TRUNCATED_SENTINEL)).toBe(true);
// Feed the already-truncated log back in with more content that forces
// another truncation.
const second = truncateLog(first, "z".repeat(1500), 1000);
expect(second.startsWith(__internal.LOG_TRUNCATED_SENTINEL)).toBe(true);
// Only one sentinel should appear in total.
const occurrences = second.split(__internal.LOG_TRUNCATED_SENTINEL).length - 1;
expect(occurrences).toBe(1);
// Most of the tail should be 'z'.
expect(second.endsWith("z".repeat(500))).toBe(true);
});
it("handles empty existing and empty addition safely", () => {
expect(truncateLog("", "", 100)).toBe("");
expect(truncateLog("", "hi", 100)).toBe("hi");
expect(truncateLog("hi", "", 100)).toBe("hi");
});
});
describe("agent-runtime — truncateError", () => {
it("leaves short strings alone", () => {
expect(truncateError("boom")).toBe("boom");
});
it("trims long strings and appends a marker", () => {
const long = "e".repeat(10_000);
const out = truncateError(long, 4096);
expect(out.length).toBeLessThanOrEqual(4096 + 32); // body + marker
expect(out.startsWith("e".repeat(4096))).toBe(true);
expect(out).toContain("truncated");
});
it("handles empty / null-ish input", () => {
expect(truncateError("")).toBe("");
// @ts-expect-error — runtime guard against null callers.
expect(truncateError(null)).toBe("");
});
});
describe("agent-runtime — runSandboxed", () => {
it("runs a successful command and captures stdout", async () => {
const res = await runSandboxed("/bin/echo", ["hello world"], {
cwd: "/tmp",
timeoutMs: 2000,
});
expect(res.exitCode).toBe(0);
expect(res.timedOut).toBe(false);
expect(res.stdout).toContain("hello world");
expect(res.stderr).toBe("");
});
it("returns a non-zero exit code for failing commands", async () => {
// `false` always exits 1.
const res = await runSandboxed("/bin/sh", ["-c", "exit 7"], {
cwd: "/tmp",
timeoutMs: 2000,
});
expect(res.exitCode).toBe(7);
expect(res.timedOut).toBe(false);
});
it("kills long-running processes on timeout", async () => {
const started = Date.now();
const res = await runSandboxed("/bin/sleep", ["10"], {
cwd: "/tmp",
timeoutMs: 200,
});
const elapsed = Date.now() - started;
expect(res.timedOut).toBe(true);
// 200ms timeout + up to 5s SIGTERM grace before SIGKILL; sleep(1) responds
// to SIGTERM immediately so in practice well under a second.
expect(elapsed).toBeLessThan(6000);
expect(res.stderr).toContain("timeout");
}, 10_000);
it("caps stdout at the configured stream cap", async () => {
// Write ~200 KB of 'a's to stdout via a tiny python-free shell loop.
// `yes` is simpler: it spams a string forever. Combined with head
// we get deterministic size without relying on Python/Node.
const res = await runSandboxed(
"/bin/sh",
["-c", "yes aaaaaaaaaa | head -c 204800"],
{ cwd: "/tmp", timeoutMs: 5000, stdoutCapBytes: 64 * 1024 }
);
expect(res.exitCode).toBe(0);
// Capped stream = 64 KB content + '\n[... truncated ...]' suffix.
expect(res.stdout.length).toBeGreaterThanOrEqual(64 * 1024);
expect(res.stdout.length).toBeLessThan(64 * 1024 + 64);
expect(res.stdout).toContain("truncated");
}, 10_000);
it("reports spawn failure without throwing", async () => {
const res = await runSandboxed(
"/nonexistent/definitely-not-a-real-binary-xyz",
[],
{ cwd: "/tmp", timeoutMs: 1000 }
);
// Bun either returns exitCode=null + stderr message, or exit!=0 — both
// are acceptable "failure surfaced" outcomes. What must NOT happen is a
// throw. Assert the call resolved and returned a result object.
expect(typeof res.timedOut).toBe("boolean");
expect(res.exitCode === null || res.exitCode !== 0).toBe(true);
});
it("passes a minimal env by default (no process.env leakage)", async () => {
// Set a secret-looking var in the parent process and ensure it does NOT
// leak into the sandbox's environment when the caller passes no env.
process.env.GLUECRON_TEST_SECRET = "leak-me-please";
const res = await runSandboxed(
"/bin/sh",
["-c", "echo \"S=${GLUECRON_TEST_SECRET:-unset}\""],
{ cwd: "/tmp", timeoutMs: 2000 }
);
delete process.env.GLUECRON_TEST_SECRET;
expect(res.exitCode).toBe(0);
expect(res.stdout).toContain("S=unset");
});
it("honours a caller-supplied env verbatim", async () => {
const res = await runSandboxed(
"/bin/sh",
["-c", "echo \"V=$FOO\""],
{
cwd: "/tmp",
timeoutMs: 2000,
env: { PATH: process.env.PATH ?? "/usr/bin:/bin", FOO: "bar" },
}
);
expect(res.exitCode).toBe(0);
expect(res.stdout).toContain("V=bar");
});
});
|