mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-10-21 08:22:20 +02:00
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The runner package is useful only when the application can start, observe, and recover a native Codex run safely. > - Existing direct adapters must keep their current execution and finalization paths. > - The application boundary therefore needs additive persistence, authorization, coordination, and recovery behind an explicit experimental adapter. > - This pull request adds that Codex-only boundary without activating generalized providers, remote environments, or the later task/SDK surfaces. ## Linked Issues or Issue Description **Subsystem affected** Shared contracts, database persistence, adapter utilities, server native-runtime services, and the experimental Paperclip Runner adapter. **Problem or motivation** The already-landed runner package has a qualified Codex path, but the application needs durable native-run state, guarded runtime selection, authenticated coordination, tool security, finalization, and recovery before the experimental adapter can be exercised safely. **Proposed solution** Add a Codex-only `paperclip_runner` application path behind the existing default-off native-runner setting. Bind native state and coordination to company/run identity, preserve persisted-run recovery, and leave every direct adapter on its existing legacy execution path. **Alternatives considered** The earlier stack boundary introduced a generalized executor and remote-environment lifecycle here. That made this PR depend on implementations in higher PRs and changed reusable sandbox behavior globally. Those pieces are now deferred together to #12592. **Roadmap alignment** ROADMAP.md does not list a conflicting native-runner integration project. This change adds the application boundary for the existing Runner architecture. ## What Changed - Added native run/result/finalization/provider-trace persistence, shared validators, and idempotent migration/replay coverage. - Added guarded Codex-only runtime selection, authenticated PRP coordination, recovery, finalization, and interaction services. - Added run/company-bound tool-gateway authorization, credential redaction, SSRF protections, and replay-safe behavior. - Added the explicit `paperclip_runner` adapter behind the default-off rollout setting. - Preserved legacy answered-question wake projection and direct-adapter execution/finalization paths. - Hardened cancellation so only owned in-memory child processes are signaled; persisted recycled PIDs/process groups are never trusted. - Retained the narrow Claude ACPX isolated-context security follow-up discovered after #12590. - Deferred the generalized executor, provider ingress, remote lifecycle, SDK/lab/eval work, release-process changes, and lockfile. ## Verification - Changed-file delta against `master`: 133 files. - GitHub Actions is the authoritative verification environment for this PR. - Full CI, security, and Greptile review will run on this lowest unmerged stack PR. - Local tests/build/typecheck were not run because this checkout is resource constrained. - Static diff/reference checks pass, and `pnpm-lock.yaml` is unchanged. ## Risks - This touches central heartbeat and agent-route code, so legacy compatibility is the primary risk. - Runtime selection remains Codex-only and explicit; direct Codex, Claude, OpenCode, process, HTTP, and plugin adapters remain on their existing paths. - Fresh native starts fail closed while the rollout flag is off; persisted native records remain readable and recoverable. - Cancellation, company/run binding, tool calls, status decisions, and completion writes are guarded or replay-safe. > For core feature work, check [ROADMAP.md](ROADMAP.md) first and discuss it in #dev before opening the PR. Feature PRs that overlap with planned core work may need to be redirected. ## Model Used OpenAI Codex, GPT-5.6, with repository tools, code execution, and parallel agent review. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either linked existing issues or described the issue in-PR following the relevant issue template - [x] I have not referenced internal or instance-local Paperclip issues or links - [x] My branch name describes the change and contains no internal Paperclip ticket id - [ ] I have run tests locally and they pass — GitHub Actions is authoritative for this resource-constrained checkout - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented risks above - [ ] All Paperclip CI and security gates are green - [ ] Greptile is 5/5 with no open actionable findings - [x] I will address all Greptile and reviewer comments before merge ## Stack - Position: 3 of 5 overall; lowest of 3 currently unmerged - Base: `master` - Previous: [#12590](https://github.com/paperclipai/paperclip/pull/12590), qualified Claude ACPX runtime — merged - Next: [#12592](https://github.com/paperclipai/paperclip/pull/12592), generalized Codex executor, task experience, and developer SDKs --------- Co-authored-by: Dev Agent <dev@paperclip.ing>
385 lines
15 KiB
TypeScript
385 lines
15 KiB
TypeScript
import fs from "node:fs/promises";
|
|
import os from "node:os";
|
|
import path from "node:path";
|
|
import { afterEach, describe, expect, it } from "vitest";
|
|
import { agentInstructionsService } from "../services/agent-instructions.js";
|
|
|
|
type TestAgent = {
|
|
id: string;
|
|
companyId: string;
|
|
name: string;
|
|
adapterConfig: Record<string, unknown>;
|
|
};
|
|
|
|
async function makeTempDir(prefix: string) {
|
|
return fs.mkdtemp(path.join(os.tmpdir(), prefix));
|
|
}
|
|
|
|
function makeAgent(adapterConfig: Record<string, unknown>): TestAgent {
|
|
return {
|
|
id: "agent-1",
|
|
companyId: "company-1",
|
|
name: "Agent 1",
|
|
adapterConfig,
|
|
};
|
|
}
|
|
|
|
describe("agent instructions service", () => {
|
|
const originalPaperclipHome = process.env.PAPERCLIP_HOME;
|
|
const originalPaperclipInstanceId = process.env.PAPERCLIP_INSTANCE_ID;
|
|
const cleanupDirs = new Set<string>();
|
|
|
|
afterEach(async () => {
|
|
if (originalPaperclipHome === undefined) delete process.env.PAPERCLIP_HOME;
|
|
else process.env.PAPERCLIP_HOME = originalPaperclipHome;
|
|
if (originalPaperclipInstanceId === undefined) delete process.env.PAPERCLIP_INSTANCE_ID;
|
|
else process.env.PAPERCLIP_INSTANCE_ID = originalPaperclipInstanceId;
|
|
|
|
await Promise.all([...cleanupDirs].map(async (dir) => {
|
|
await fs.rm(dir, { recursive: true, force: true });
|
|
cleanupDirs.delete(dir);
|
|
}));
|
|
});
|
|
|
|
it("copies the existing bundle into the managed root when switching to managed mode", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-home-");
|
|
const externalRoot = await makeTempDir("paperclip-agent-instructions-external-");
|
|
cleanupDirs.add(paperclipHome);
|
|
cleanupDirs.add(externalRoot);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
await fs.writeFile(path.join(externalRoot, "AGENTS.md"), "# External Agent\n", "utf8");
|
|
await fs.mkdir(path.join(externalRoot, "docs"), { recursive: true });
|
|
await fs.writeFile(path.join(externalRoot, "docs", "TOOLS.md"), "## Tools\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "external",
|
|
instructionsRootPath: externalRoot,
|
|
instructionsEntryFile: "AGENTS.md",
|
|
instructionsFilePath: path.join(externalRoot, "AGENTS.md"),
|
|
});
|
|
|
|
const result = await svc.updateBundle(agent, { mode: "managed" });
|
|
|
|
expect(result.bundle.mode).toBe("managed");
|
|
expect(result.bundle.managedRootPath).toBe(
|
|
path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
),
|
|
);
|
|
expect(result.bundle.files.map((file) => file.path)).toEqual(["AGENTS.md", "docs/TOOLS.md"]);
|
|
await expect(fs.readFile(path.join(result.bundle.managedRootPath, "AGENTS.md"), "utf8")).resolves.toBe("# External Agent\n");
|
|
await expect(fs.readFile(path.join(result.bundle.managedRootPath, "docs", "TOOLS.md"), "utf8")).resolves.toBe("## Tools\n");
|
|
});
|
|
|
|
it("creates the target entry file when switching to a new external root", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-home-");
|
|
const managedRoot = path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
);
|
|
const externalRoot = await makeTempDir("paperclip-agent-instructions-new-external-");
|
|
cleanupDirs.add(paperclipHome);
|
|
cleanupDirs.add(externalRoot);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
await fs.mkdir(managedRoot, { recursive: true });
|
|
await fs.writeFile(path.join(managedRoot, "AGENTS.md"), "# Managed Agent\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "managed",
|
|
instructionsRootPath: managedRoot,
|
|
instructionsEntryFile: "AGENTS.md",
|
|
instructionsFilePath: path.join(managedRoot, "AGENTS.md"),
|
|
});
|
|
|
|
const result = await svc.updateBundle(agent, {
|
|
mode: "external",
|
|
rootPath: externalRoot,
|
|
entryFile: "docs/AGENTS.md",
|
|
});
|
|
|
|
expect(result.bundle.mode).toBe("external");
|
|
expect(result.bundle.rootPath).toBe(externalRoot);
|
|
await expect(fs.readFile(path.join(externalRoot, "docs", "AGENTS.md"), "utf8")).resolves.toBe("# Managed Agent\n");
|
|
});
|
|
|
|
it("filters junk files, dependency bundles, and python caches from bundle listings and exports", async () => {
|
|
const externalRoot = await makeTempDir("paperclip-agent-instructions-ignore-");
|
|
cleanupDirs.add(externalRoot);
|
|
|
|
await fs.writeFile(path.join(externalRoot, "AGENTS.md"), "# External Agent\n", "utf8");
|
|
await fs.writeFile(path.join(externalRoot, ".gitignore"), "node_modules/\n", "utf8");
|
|
await fs.writeFile(path.join(externalRoot, ".DS_Store"), "junk", "utf8");
|
|
await fs.mkdir(path.join(externalRoot, "docs"), { recursive: true });
|
|
await fs.writeFile(path.join(externalRoot, "docs", "TOOLS.md"), "## Tools\n", "utf8");
|
|
await fs.writeFile(path.join(externalRoot, "docs", "module.pyc"), "compiled", "utf8");
|
|
await fs.writeFile(path.join(externalRoot, "docs", "._TOOLS.md"), "appledouble", "utf8");
|
|
await fs.mkdir(path.join(externalRoot, "node_modules", "pkg"), { recursive: true });
|
|
await fs.writeFile(path.join(externalRoot, "node_modules", "pkg", "index.js"), "export {};\n", "utf8");
|
|
await fs.mkdir(path.join(externalRoot, "python", "__pycache__"), { recursive: true });
|
|
await fs.writeFile(
|
|
path.join(externalRoot, "python", "__pycache__", "module.cpython-313.pyc"),
|
|
"compiled",
|
|
"utf8",
|
|
);
|
|
await fs.mkdir(path.join(externalRoot, ".pytest_cache"), { recursive: true });
|
|
await fs.writeFile(path.join(externalRoot, ".pytest_cache", "README.md"), "cache", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "external",
|
|
instructionsRootPath: externalRoot,
|
|
instructionsEntryFile: "AGENTS.md",
|
|
instructionsFilePath: path.join(externalRoot, "AGENTS.md"),
|
|
});
|
|
|
|
const bundle = await svc.getBundle(agent);
|
|
const exported = await svc.exportFiles(agent);
|
|
|
|
expect(bundle.files.map((file) => file.path)).toEqual([".gitignore", "AGENTS.md", "docs/TOOLS.md"]);
|
|
expect(Object.keys(exported.files).sort((left, right) => left.localeCompare(right))).toEqual([
|
|
".gitignore",
|
|
"AGENTS.md",
|
|
"docs/TOOLS.md",
|
|
]);
|
|
});
|
|
|
|
it.skipIf(process.platform === "win32")("rejects instruction symlinks for immutable runner snapshots without changing legacy exports", async () => {
|
|
const externalRoot = await makeTempDir("paperclip-agent-instructions-symlink-");
|
|
const outsideRoot = await makeTempDir("paperclip-agent-instructions-outside-");
|
|
cleanupDirs.add(externalRoot);
|
|
cleanupDirs.add(outsideRoot);
|
|
await fs.writeFile(path.join(externalRoot, "AGENTS.md"), "Read sibling.md\n", "utf8");
|
|
await fs.writeFile(path.join(outsideRoot, "secret.md"), "must not enter the bundle\n", "utf8");
|
|
await fs.symlink(path.join(outsideRoot, "secret.md"), path.join(externalRoot, "sibling.md"));
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "external",
|
|
instructionsRootPath: externalRoot,
|
|
instructionsEntryFile: "AGENTS.md",
|
|
instructionsFilePath: path.join(externalRoot, "AGENTS.md"),
|
|
});
|
|
const svc = agentInstructionsService();
|
|
|
|
await expect(svc.exportFiles(agent)).resolves.toMatchObject({
|
|
files: { "AGENTS.md": "Read sibling.md\n" },
|
|
});
|
|
await expect(svc.exportFiles(agent, { rejectSymlinks: true }))
|
|
.rejects.toThrow("Instructions bundle may not contain symlinks: sibling.md");
|
|
});
|
|
|
|
it("recovers a managed bundle from disk when bundle config metadata is missing", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-recover-");
|
|
cleanupDirs.add(paperclipHome);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
const managedRoot = path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
);
|
|
await fs.mkdir(managedRoot, { recursive: true });
|
|
await fs.writeFile(path.join(managedRoot, "AGENTS.md"), "# Recovered Agent\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({});
|
|
|
|
const bundle = await svc.getBundle(agent);
|
|
const exported = await svc.exportFiles(agent);
|
|
|
|
expect(bundle.mode).toBe("managed");
|
|
expect(bundle.rootPath).toBe(managedRoot);
|
|
expect(bundle.files.map((file) => file.path)).toEqual(["AGENTS.md"]);
|
|
expect(exported.files).toEqual({ "AGENTS.md": "# Recovered Agent\n" });
|
|
});
|
|
|
|
it("prefers the managed bundle on disk when managed metadata points at a stale root", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-stale-managed-");
|
|
const staleRoot = await makeTempDir("paperclip-agent-instructions-stale-root-");
|
|
cleanupDirs.add(paperclipHome);
|
|
cleanupDirs.add(staleRoot);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
const managedRoot = path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
);
|
|
await fs.mkdir(managedRoot, { recursive: true });
|
|
await fs.writeFile(path.join(managedRoot, "AGENTS.md"), "# Managed Agent\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "managed",
|
|
instructionsRootPath: staleRoot,
|
|
instructionsEntryFile: "docs/MISSING.md",
|
|
instructionsFilePath: path.join(staleRoot, "docs", "MISSING.md"),
|
|
});
|
|
|
|
const bundle = await svc.getBundle(agent);
|
|
const exported = await svc.exportFiles(agent);
|
|
|
|
expect(bundle.mode).toBe("managed");
|
|
expect(bundle.rootPath).toBe(managedRoot);
|
|
expect(bundle.entryFile).toBe("AGENTS.md");
|
|
expect(bundle.files.map((file) => file.path)).toEqual(["AGENTS.md"]);
|
|
expect(bundle.warnings).toEqual([
|
|
`Recovered managed instructions from disk at ${managedRoot}; ignoring stale configured root ${staleRoot}.`,
|
|
"Recovered managed instructions entry file from disk as AGENTS.md; previous entry docs/MISSING.md was missing.",
|
|
]);
|
|
expect(exported.files).toEqual({ "AGENTS.md": "# Managed Agent\n" });
|
|
});
|
|
|
|
it("heals stale managed metadata when writing bundle files", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-heal-write-");
|
|
const staleRoot = await makeTempDir("paperclip-agent-instructions-heal-write-stale-");
|
|
cleanupDirs.add(paperclipHome);
|
|
cleanupDirs.add(staleRoot);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
const managedRoot = path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
);
|
|
await fs.mkdir(path.join(managedRoot, "docs"), { recursive: true });
|
|
await fs.writeFile(path.join(managedRoot, "AGENTS.md"), "# Managed Agent\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "managed",
|
|
instructionsRootPath: staleRoot,
|
|
instructionsEntryFile: "docs/MISSING.md",
|
|
instructionsFilePath: path.join(staleRoot, "docs", "MISSING.md"),
|
|
});
|
|
|
|
const result = await svc.writeFile(agent, "docs/TOOLS.md", "## Tools\n");
|
|
|
|
expect(result.adapterConfig).toMatchObject({
|
|
instructionsBundleMode: "managed",
|
|
instructionsRootPath: managedRoot,
|
|
instructionsEntryFile: "AGENTS.md",
|
|
instructionsFilePath: path.join(managedRoot, "AGENTS.md"),
|
|
});
|
|
await expect(fs.readFile(path.join(managedRoot, "docs", "TOOLS.md"), "utf8")).resolves.toBe("## Tools\n");
|
|
});
|
|
|
|
it("heals stale managed metadata when deleting bundle files", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-heal-delete-");
|
|
const staleRoot = await makeTempDir("paperclip-agent-instructions-heal-delete-stale-");
|
|
cleanupDirs.add(paperclipHome);
|
|
cleanupDirs.add(staleRoot);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
const managedRoot = path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
);
|
|
await fs.mkdir(path.join(managedRoot, "docs"), { recursive: true });
|
|
await fs.writeFile(path.join(managedRoot, "AGENTS.md"), "# Managed Agent\n", "utf8");
|
|
await fs.writeFile(path.join(managedRoot, "docs", "TOOLS.md"), "## Tools\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsBundleMode: "managed",
|
|
instructionsRootPath: staleRoot,
|
|
instructionsEntryFile: "docs/MISSING.md",
|
|
instructionsFilePath: path.join(staleRoot, "docs", "MISSING.md"),
|
|
});
|
|
|
|
const result = await svc.deleteFile(agent, "docs/TOOLS.md");
|
|
|
|
expect(result.adapterConfig).toMatchObject({
|
|
instructionsBundleMode: "managed",
|
|
instructionsRootPath: managedRoot,
|
|
instructionsEntryFile: "AGENTS.md",
|
|
instructionsFilePath: path.join(managedRoot, "AGENTS.md"),
|
|
});
|
|
await expect(fs.stat(path.join(managedRoot, "docs", "TOOLS.md"))).rejects.toThrow();
|
|
expect(result.bundle.files.map((file) => file.path)).toEqual(["AGENTS.md"]);
|
|
});
|
|
|
|
it("recovers the managed bundle when stale root metadata is present but mode is missing", async () => {
|
|
const paperclipHome = await makeTempDir("paperclip-agent-instructions-partial-managed-");
|
|
const staleRoot = await makeTempDir("paperclip-agent-instructions-partial-root-");
|
|
cleanupDirs.add(paperclipHome);
|
|
cleanupDirs.add(staleRoot);
|
|
process.env.PAPERCLIP_HOME = paperclipHome;
|
|
process.env.PAPERCLIP_INSTANCE_ID = "test-instance";
|
|
|
|
const managedRoot = path.join(
|
|
paperclipHome,
|
|
"instances",
|
|
"test-instance",
|
|
"companies",
|
|
"company-1",
|
|
"agents",
|
|
"agent-1",
|
|
"instructions",
|
|
);
|
|
await fs.mkdir(managedRoot, { recursive: true });
|
|
await fs.writeFile(path.join(managedRoot, "AGENTS.md"), "# Managed Agent\n", "utf8");
|
|
|
|
const svc = agentInstructionsService();
|
|
const agent = makeAgent({
|
|
instructionsRootPath: staleRoot,
|
|
instructionsEntryFile: "docs/MISSING.md",
|
|
});
|
|
|
|
const bundle = await svc.getBundle(agent);
|
|
const exported = await svc.exportFiles(agent);
|
|
|
|
expect(bundle.mode).toBe("managed");
|
|
expect(bundle.rootPath).toBe(managedRoot);
|
|
expect(bundle.entryFile).toBe("AGENTS.md");
|
|
expect(bundle.files.map((file) => file.path)).toEqual(["AGENTS.md"]);
|
|
expect(bundle.warnings).toEqual([
|
|
`Recovered managed instructions from disk at ${managedRoot}; ignoring stale configured root ${staleRoot}.`,
|
|
"Recovered managed instructions entry file from disk as AGENTS.md; previous entry docs/MISSING.md was missing.",
|
|
]);
|
|
expect(exported.files).toEqual({ "AGENTS.md": "# Managed Agent\n" });
|
|
});
|
|
});
|