evOo-local/infra/sandboxes/supervisor/src/github-broker.ts

import { spawn } from "node:child_process";
import { statSync } from "node:fs";
import path from "node:path";
import {
  evaluateGitShellPolicy,
  ghBrokerArgv,
  gitNetworkArgv,
  gitRemoteMatchesAllowlist,
  normalizeGitRemoteUrl,
  resolveSupervisorToken,
} from "@rakazo/core";
import { verifyGithubBrokerRunToken } from "@rakazo/core/node/github-broker-token";
import type { Hono } from "hono";
import { z } from "zod";
import { hasValidBearerToken } from "./supervisor-logic.js";

export type GithubBrokerSession = {
  runId: string;
  workspacePath: string;
  allowedRemoteUrl: string | null;
  defaultBranch: string | null;
  githubToken: string;
  /** Where workspacePath is mounted inside the computer. */
  containerRoot: string;
  pushApproved: boolean;
};

const DEFAULT_CONTAINER_ROOT = "/home/rakazo";

// Plain git ignores GITHUB_TOKEN, so HTTPS remotes get the token from a one-off credential helper.
const GIT_CREDENTIAL_ARGS = [
  "-c",
  "credential.helper=",
  "-c",
  'credential.https://github.com.helper=!f() { test "$1" = get && echo username=x-access-token && echo "password=$GITHUB_TOKEN"; }; f',
];

/** Maps a path inside the computer onto the host workspace; anything outside the mount stays unmapped. */
export function hostPathFor(session: GithubBrokerSession, containerPath: string): string | null {
  const root = session.containerRoot.replace(/\/$/, "");
  if (containerPath !== root && !containerPath.startsWith(`${root}/`)) return null;
  const resolved = path.resolve(session.workspacePath, path.relative(root, containerPath));
  const inside = path.relative(session.workspacePath, resolved);
  return inside.startsWith("..") || path.isAbsolute(inside) ? null : resolved;
}

/** GitHub remotes always go over HTTPS with the broker's token: drop pasted credentials, turn SSH into HTTPS. */
export function githubHttpsRemote(arg: string): string {
  const ssh = /^(?:ssh:\/\/)?git@github\.com[:/](.+)$/.exec(arg);
  if (ssh) return `https://github.com/${ssh[1]}`;
  return arg.replace(/^https:\/\/[^/@]+@github\.com\//, "https://github.com/");
}

const GITHUB_AUTH_FAILURE =
  /Invalid username or token|Authentication failed|could not read Username/;

/** GitHub rejecting the stored token means the connection needs renewing, not a token from the user. */
export function explainGitFailure(stderr: string): string {
  if (!GITHUB_AUTH_FAILURE.test(stderr)) return stderr;
  return `${stderr.trimEnd()}\nGitHub rejected the token evOo stores for this user. Tell the user to reconnect GitHub under Settings, Integrations. Do not ask for a token or set one up yourself.\n`;
}

/** Host argv and cwd for a git call made in the computer, or an error when it leaves the workspace. */
export function hostGitInvocation(
  session: GithubBrokerSession,
  argv: string[],
  containerCwd: string | undefined,
): { argv: string[]; cwd: string } | { error: string } {
  const cwd = containerCwd ? hostPathFor(session, containerCwd) : session.workspacePath;
  if (!cwd) return { error: `Run git inside ${session.containerRoot}.` };
  const mapped: string[] = [];
  for (const part of argv.slice(1)) {
    if (!part.startsWith("/")) {
      mapped.push(githubHttpsRemote(part));
      continue;
    }
    const hostPart = hostPathFor(session, part);
    if (!hostPart) return { error: `Use a path inside ${session.containerRoot}.` };
    mapped.push(hostPart);
  }
  return { argv: ["/usr/bin/git", ...GIT_CREDENTIAL_ARGS, ...mapped], cwd };
}

/** A root supervisor runs git as the workspace owner, so clones stay writable by the computer. */
export function gitProcessIdentity(
  workspacePath: string,
  uid = process.getuid?.(),
): { uid: number; gid: number } | undefined {
  if (uid !== 0) return undefined;
  try {
    const stat = statSync(workspacePath);
    return stat.uid === 0 ? undefined : { uid: stat.uid, gid: stat.gid };
  } catch {
    return undefined;
  }
}

export type BrokerCommandResult = { stdout: string; stderr: string; code: number };

export type BrokerCommandInput = {
  argv: string[];
  cwd: string;
  env: Record<string, string>;
  workspacePath: string;
};

/** Runs git or gh for a workspace this process cannot see, e.g. a project folder outside a container's mounts. */
export type BrokerOutsideRunner = {
  sees(target: string): boolean;
  run(input: BrokerCommandInput): Promise<BrokerCommandResult>;
};

let runOutside: BrokerOutsideRunner | undefined;

const sessions = new Map<string, GithubBrokerSession>();

export function githubBrokerSessionsForTests() {
  return sessions;
}

function brokerSecret(): string {
  return resolveSupervisorToken(process.env);
}

function runIdFromRunToken(authorization: string | undefined): string | null {
  if (!authorization?.startsWith("Bearer ")) return null;
  const token = authorization.slice("Bearer ".length).trim();
  for (const runId of sessions.keys()) {
    if (verifyGithubBrokerRunToken(token, brokerSecret(), runId)) return runId;
  }
  return null;
}

function sessionForRequest(authorization: string | undefined): GithubBrokerSession | null {
  const runId = runIdFromRunToken(authorization);
  if (!runId) return null;
  return sessions.get(runId) ?? null;
}

async function runHostGit(
  session: GithubBrokerSession,
  argv: string[],
  cwd = session.workspacePath,
): Promise<BrokerCommandResult> {
  const env = {
    HOME: session.workspacePath,
    GIT_TERMINAL_PROMPT: "0",
    GH_TOKEN: session.githubToken,
    GITHUB_TOKEN: session.githubToken,
  };
  if (runOutside && !runOutside.sees(cwd)) {
    return runOutside.run({ argv, cwd, env, workspacePath: session.workspacePath });
  }
  return new Promise((resolve) => {
    const child = spawn(argv[0]!, argv.slice(1), {
      cwd,
      ...gitProcessIdentity(session.workspacePath),
      env: { ...process.env, ...env },
    });
    let stdout = "";
    let stderr = "";
    child.stdout?.on("data", (chunk: Buffer) => {
      stdout += chunk.toString("utf8");
    });
    child.stderr?.on("data", (chunk: Buffer) => {
      stderr += chunk.toString("utf8");
    });
    child.on("close", (code) => resolve({ stdout, stderr, code: code ?? 1 }));
    child.on("error", (error) => resolve({ stdout: "", stderr: error.message, code: 1 }));
  });
}

async function readRemoteUrl(
  session: GithubBrokerSession,
  remoteName: string,
  cwd = session.workspacePath,
): Promise<string | null> {
  const result = await runHostGit(
    session,
    ["/usr/bin/git", "remote", "get-url", "--", remoteName],
    cwd,
  );
  if (result.code !== 0) return null;
  return result.stdout.trim() || null;
}

export async function assertRemoteAllowed(
  session: GithubBrokerSession,
  argv: string[],
  cwd = session.workspacePath,
) {
  if (!session.allowedRemoteUrl) return;
  const remoteArg = argv.find((part, index) => index > 1 && !part.startsWith("-"));
  const candidate = remoteArg
    ? normalizeGitRemoteUrl(remoteArg)
      ? remoteArg
      : ((await readRemoteUrl(session, remoteArg, cwd)) ?? remoteArg)
    : await readRemoteUrl(session, "origin", cwd);
  if (candidate && !gitRemoteMatchesAllowlist(candidate, session.allowedRemoteUrl)) {
    throw new Error("Push and fetch are limited to this space's linked repository remote.");
  }
}

async function resolveDefaultBranch(
  session: GithubBrokerSession,
  cwd = session.workspacePath,
): Promise<string | null> {
  if (session.defaultBranch) return session.defaultBranch;
  const sym = await runHostGit(
    session,
    ["/usr/bin/git", "symbolic-ref", "refs/remotes/origin/HEAD"],
    cwd,
  );
  if (sym.code !== 0) return null;
  const ref = sym.stdout.trim();
  const match = /^refs\/remotes\/origin\/(.+)$/.exec(ref);
  return match?.[1]?.trim() || null;
}

export function mountGithubBrokerRoutes(
  app: Hono,
  supervisorToken: string,
  outside?: BrokerOutsideRunner,
) {
  if (outside) runOutside = outside;
  app.use("/internal/github-broker/*", async (c, next) => {
    const pathname = c.req.path;
    if (pathname.startsWith("/internal/github-broker/sessions")) {
      if (!hasValidBearerToken(c.req.header("authorization"), supervisorToken)) {
        return c.json({ error: "unauthorized" }, 401);
      }
      await next();
      return;
    }
    if (!sessionForRequest(c.req.header("authorization"))) {
      return c.json({ error: "unauthorized" }, 401);
    }
    await next();
  });

  app.post("/internal/github-broker/sessions", async (c) => {
    const body = z
      .object({
        runId: z.string().min(1),
        workspacePath: z.string().min(1),
        allowedRemoteUrl: z.string().nullable(),
        defaultBranch: z.string().nullable(),
        githubToken: z.string().min(1),
        containerRoot: z.string().min(1).optional(),
        pushApproved: z.boolean().optional(),
      })
      .parse(await c.req.json());
    const workspacePath = path.resolve(body.workspacePath);
    sessions.set(body.runId, {
      runId: body.runId,
      workspacePath,
      allowedRemoteUrl: body.allowedRemoteUrl,
      defaultBranch: body.defaultBranch,
      githubToken: body.githubToken,
      containerRoot: body.containerRoot ?? DEFAULT_CONTAINER_ROOT,
      pushApproved: body.pushApproved ?? false,
    });
    return c.json({ ok: true });
  });

  app.patch("/internal/github-broker/sessions/:runId", async (c) => {
    const runId = c.req.param("runId");
    const body = z.object({ pushApproved: z.boolean() }).parse(await c.req.json());
    const session = sessions.get(runId);
    if (!session) return c.json({ error: "not found" }, 404);
    session.pushApproved = body.pushApproved;
    return c.json({ ok: true });
  });

  app.delete("/internal/github-broker/sessions/:runId", async (c) => {
    sessions.delete(c.req.param("runId"));
    return c.json({ ok: true });
  });

  app.post("/internal/github-broker/v1/git", async (c) => {
    const session = sessionForRequest(c.req.header("authorization"));
    if (!session) return c.json({ error: "unauthorized" }, 401);
    const body = z
      .object({ argv: z.array(z.string()).min(1), cwd: z.string().optional() })
      .parse(await c.req.json());
    const argv = body.argv[0] === "git" ? body.argv : ["git", ...body.argv];
    if (!gitNetworkArgv(argv)) {
      return c.json({ error: "unsupported git invocation" }, 400);
    }
    const invocation = hostGitInvocation(session, argv, body.cwd);
    if ("error" in invocation) return c.json({ stdout: "", stderr: invocation.error, code: 1 });
    const command = argv.join(" ");
    const defaultBranch = await resolveDefaultBranch(session, invocation.cwd);
    const policy = evaluateGitShellPolicy({
      command,
      defaultBranch,
      gitPushApprovedThisRun: session.pushApproved,
    });
    if (policy.action === "block") {
      return c.json({ stdout: "", stderr: policy.reason, code: 1 });
    }
    if (policy.action === "require_approval") {
      return c.json({ stdout: "", stderr: policy.reason, code: 1 });
    }
    try {
      await assertRemoteAllowed(session, argv, invocation.cwd);
    } catch (error) {
      const message = error instanceof Error ? error.message : String(error);
      return c.json({ stdout: "", stderr: message, code: 1 });
    }
    const result = await runHostGit(session, invocation.argv, invocation.cwd);
    return c.json({ ...result, stderr: explainGitFailure(result.stderr) });
  });

  app.post("/internal/github-broker/v1/gh", async (c) => {
    const session = sessionForRequest(c.req.header("authorization"));
    if (!session) return c.json({ error: "unauthorized" }, 401);
    const body = z.object({ argv: z.array(z.string()).min(1) }).parse(await c.req.json());
    const argv = body.argv[0] === "gh" ? body.argv : ["gh", ...body.argv];
    if (!ghBrokerArgv(argv)) {
      return c.json({ error: "unsupported gh invocation" }, 400);
    }
    const defaultBranch = await resolveDefaultBranch(session);
    const policy = evaluateGitShellPolicy({
      command: argv.join(" "),
      defaultBranch,
      gitPushApprovedThisRun: session.pushApproved,
    });
    if (policy.action === "block" || policy.action === "require_approval") {
      const reason = policy.reason;
      return c.json({ stdout: "", stderr: reason, code: 1 });
    }
    const result = await runHostGit(session, ["/usr/bin/gh", ...argv.slice(1)]);
    return c.json(result);
  });
}