fix(wrap): set xAI upstream for grok-build proxy (#2772)

## Description

`headroom wrap grok-build` injected the client hop into
`~/.grok/config.toml` but started the local proxy **without** setting
the OpenAI-compatible upstream to xAI. The proxy defaulted to
`api.openai.com`, so Grok session auth returned **401** on every chat
completion even though compression still ran.

`wrap grok` already passes `openai_api_url` → xAI. This PR aligns `wrap
grok-build` and the Grok-only persistent `install` path on the shared
`DEFAULT_API_URL` (`https://api.x.ai`).

Closes # (none — discovered in live Grok Build pilot)

## Type of Change

- [x] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to change)
- [ ] Documentation update
- [ ] Performance improvement
- [ ] Code refactoring (no functional changes)

## Changes Made

- Pass `openai_api_url=_GROK_DEFAULT_API_URL` into
`_run_proxy_only_watcher` from `wrap grok-build`
- Use shared `DEFAULT_API_URL` from `wrap grok` (no hard-coded string
drift)
- Print proxy upstream in Grok Build setup lines
- Persistent install: when targets are Grok-only, set
`OPENAI_TARGET_API_URL` + `--openai-api-url` (skip when
Codex/Copilot/Aider/OpenCode share the proxy; explicit env still wins)
- Regression tests for wrap kwargs, setup lines, and install planner

## Testing

- [x] Unit tests pass (`pytest` targeted suite)
- [ ] Linting passes (`ruff check .`) — not run in this environment (no
native editable build)
- [ ] Type checking passes (`mypy headroom`) — not run
- [x] New tests added for new functionality
- [x] Manual testing performed

### Test Output

```text
$ PYTHONPATH=$PWD python -m pytest \
    tests/test_cli/test_wrap_bridge.py::test_wrap_grok_build_passes_xai_openai_api_url \
    tests/test_cli/test_wrap_bridge.py::test_wrap_grok_build_uses_actual_proxy_port \
    tests/test_install/test_planner.py::test_build_manifest_grok_build_only_sets_xai_upstream \
    tests/test_install/test_planner.py::test_build_manifest_grok_with_codex_does_not_force_xai \
    tests/test_install/test_planner.py::test_build_manifest_extra_env_wins_over_grok_xai_default \
    tests/test_provider_grok_build.py::test_grok_build_setup_lines_include_proxy_url -q
......
6 passed in 0.33s
```

## Real Behavior Proof

- Environment: macOS (darwin), Headroom 0.33.0 via `uv tool install
"headroom-ai[proxy,mcp,code]==0.33.0"`, Grok Build CLI, models
`grok-build` and `grok-4.5`, proxy on `127.0.0.1:8787`, upstream must be
xAI
- Exact command / steps: (1) Before: stock `headroom wrap grok-build`
then `grok -m grok-build` one-shot prompt. (2) After: same wrap path
with this branch (`openai_api_url=DEFAULT_API_URL` into
`_run_proxy_only_watcher`) then `grok -m grok-build -p
'…HEADROOM_XAI_OK…'`. Also exercised `grok-4.5` via `[model."grok-4.5"]
base_url` → same proxy.
- Observed result: Before — proxy log outbound `api.openai.com` → HTTP
401; client failed while local compression still ran. After — setup line
prints Proxy upstream `https://api.x.ai`; proxy log `POST
https://api.x.ai/v1/chat/completions` (and `/v1/responses` for grok-4.5)
→ status=200; dashboard shows 0 failed requests and accumulating token
savings on live traffic.
- Not tested: full `uv run` editable/maturin native build on this host;
multi-tool install matrix beyond planner unit tests; Windows; ruff/mypy
full tree

## Review Readiness

- [x] I have performed a self-review
- [x] This PR is ready for human review

## Checklist

- [x] My code follows the project's style guidelines
- [x] I have performed a self-review of my code
- [x] I commented my code, particularly in hard-to-understand areas
- [ ] I made corresponding changes to the documentation (CLI help text /
setup lines only)
- [x] My changes generate no new warnings
- [x] I added tests that prove my fix is effective or that my feature
works
- [x] New and existing unit tests pass locally with my changes
- [x] I did **not** edit `CHANGELOG.md` — generated by release-please
from Conventional Commit PR title (a CI guard enforces this)

## Additional Notes

- Intentional non-goal: changing default model, savings %, or Grok Build
context-tool defaults
- Mixed-target install (e.g. `grok_build` + `codex`) does **not** force
xAI — operator must set upstream explicitly if they share one proxy
- Related live routing: manual `[model."grok-4.5"] base_url` through the
same proxy works once upstream is xAI (`/v1/responses`)

---------

Co-authored-by: Grok 4.5 <noreply@x.ai>
Co-authored-by: Nestor G Pestelos Jr <ngpestelos@me.com>
Co-authored-by: JD Davis <jd@jds-macbook-air.tail2a279.ts.net>
This commit is contained in:
Nestor G Pestelos Jr
2026-08-17 06:04:59 +08:00
committed by GitHub
parent 6d87825f62
commit c8310819a4
6 changed files with 110 additions and 5 deletions
+13 -5
View File
@@ -136,7 +136,12 @@ from headroom.providers.copilot import (
validate_configuration as _validate_copilot_configuration,
)
from headroom.providers.cursor import render_setup_lines as _render_cursor_setup_lines
from headroom.providers.grok import build_launch_env as _build_grok_launch_env
from headroom.providers.grok import (
DEFAULT_API_URL as _GROK_DEFAULT_API_URL,
)
from headroom.providers.grok import (
build_launch_env as _build_grok_launch_env,
)
from headroom.providers.grok_build import render_setup_lines as _render_grok_build_setup_lines
from headroom.providers.grok_build.config import (
inject_grok_provider_config,
@@ -6417,7 +6422,7 @@ def grok(
backend=backend,
anyllm_provider=anyllm_provider,
region=region,
openai_api_url="https://api.x.ai",
openai_api_url=_GROK_DEFAULT_API_URL,
)
@@ -6507,9 +6512,9 @@ def grok_build(
\b
Grok Build reads model endpoints from ``~/.grok/config.toml``. This
command starts the proxy, optionally sets up the selected CLI context
tool, injects a Headroom-managed ``[model.grok-build]`` override, and
prints next steps.
command starts the proxy (upstream ``https://api.x.ai``, same as
``wrap grok``), injects a Headroom-managed ``[model.grok-build]``
override, and prints next steps.
\b
Example:
@@ -6536,6 +6541,8 @@ def grok_build(
for line in _render_grok_build_setup_lines(actual_port, project=project):
click.echo(line)
# Client hop is local proxy via config.toml; upstream must be xAI (not
# the OpenAI default). Omitting this caused 401s with Grok auth headers.
_run_proxy_only_watcher(
agent_label="grok-build",
port=port,
@@ -6544,6 +6551,7 @@ def grok_build(
memory=memory,
agent_type="grok_build",
print_setup_lines=_print_grok_build_setup,
openai_api_url=_GROK_DEFAULT_API_URL,
)
+17
View File
@@ -9,6 +9,7 @@ from collections.abc import Iterable
import click
from headroom import paths as _paths
from headroom.providers.grok.runtime import DEFAULT_API_URL as _GROK_DEFAULT_API_URL
from headroom.providers.install_registry import build_install_target_envs
from headroom.rollout import RolloutChannel
@@ -177,6 +178,19 @@ def build_manifest(
base_env["HEADROOM_TELEMETRY"] = "on" if telemetry_enabled else "off"
if memory_enabled:
base_env["HEADROOM_MEMORY_ENABLED"] = "1"
# Grok / Grok Build need proxy upstream = xAI. Only auto-set when no other
# OpenAI-compatible tools share this proxy (those may need api.openai.com /
# Copilot). Explicit OPENAI_TARGET_API_URL in extra_env still wins below.
_openai_native = {
ToolTarget.CODEX.value,
ToolTarget.COPILOT.value,
ToolTarget.AIDER.value,
ToolTarget.OPENCODE.value,
}
_grok_targets = {ToolTarget.GROK.value, ToolTarget.GROK_BUILD.value}
target_set = set(resolved_targets)
if target_set & _grok_targets and not (target_set & _openai_native):
base_env.setdefault("OPENAI_TARGET_API_URL", _GROK_DEFAULT_API_URL)
# Applied last so explicit --env overrides win over the auto-derived
# defaults above (e.g. a custom HEADROOM_WORKSPACE_DIR).
if extra_env:
@@ -241,6 +255,9 @@ def build_manifest(
proxy_args.extend(["--protect-tool-results", protect_tool_results])
if bedrock_profile:
proxy_args.extend(["--bedrock-profile", bedrock_profile])
openai_target = base_env.get("OPENAI_TARGET_API_URL")
if openai_target:
proxy_args.extend(["--openai-api-url", openai_target])
container_name = f"headroom-{normalized_profile}"
return DeploymentManifest(
+3
View File
@@ -4,6 +4,7 @@ from __future__ import annotations
from dataclasses import dataclass
from headroom.providers.grok.runtime import DEFAULT_API_URL
from headroom.proxy.project_context import with_project_prefix
@@ -41,6 +42,8 @@ def render_setup_lines(port: int, project: str | None = None) -> list[str]:
" [model.grok-build]",
f' base_url = "{target.base_url}"',
"",
f" Proxy upstream (OpenAI-compatible): {DEFAULT_API_URL}",
"",
" Start Grok Build in this project directory:",
" grok",
"",
+34
View File
@@ -74,6 +74,40 @@ def test_wrap_grok_build_uses_actual_proxy_port(monkeypatch, tmp_path: Path) ->
assert "http://127.0.0.1:8787/" not in result.output
def test_wrap_grok_build_passes_xai_openai_api_url(monkeypatch, tmp_path: Path) -> None:
"""Grok Build must set proxy upstream to xAI (same as wrap grok).
Without openai_api_url, the proxy defaults to api.openai.com and Grok
session auth returns 401 on every chat completion.
"""
from headroom.providers.grok import DEFAULT_API_URL
_set_test_home(monkeypatch, tmp_path)
runner = CliRunner()
captured: dict = {}
def fake_watcher(**kwargs) -> None:
captured.update(kwargs)
kwargs["print_setup_lines"](kwargs["port"])
monkeypatch.setattr("headroom.cli.wrap._run_proxy_only_watcher", fake_watcher)
result = runner.invoke(main, ["wrap", "grok-build", "--port", "8787"])
assert result.exit_code == 0, result.output
assert captured.get("openai_api_url") == DEFAULT_API_URL
# Equality on the constant (not substring containment) keeps CodeQL
# incomplete-url-substring-sanitization quiet while pinning the host.
assert DEFAULT_API_URL == "https://api.x.ai"
expected_upstream = f" Proxy upstream (OpenAI-compatible): {DEFAULT_API_URL}"
upstream_lines = [
line
for line in result.output.splitlines()
if line.startswith(" Proxy upstream (OpenAI-compatible): ")
]
assert upstream_lines == [expected_upstream]
def test_wrap_rejects_retired_context_tool_flag(monkeypatch, tmp_path: Path) -> None:
"""A surviving --context-tool must fail loudly, not be silently ignored.
+35
View File
@@ -330,3 +330,38 @@ def test_build_manifest_extra_env_overrides_derived_defaults() -> None:
# telemetry_enabled=False in _base_manifest_kwargs would normally set "off";
# an explicit --env must win.
assert manifest.base_env["HEADROOM_TELEMETRY"] == "on"
def test_build_manifest_grok_build_only_sets_xai_upstream() -> None:
"""Persistent install for Grok Build alone must route proxy upstream to xAI."""
from headroom.providers.grok import DEFAULT_API_URL
manifest = build_manifest(**_base_manifest_kwargs(targets=["grok_build"], backend="openai"))
assert manifest.base_env.get("OPENAI_TARGET_API_URL") == DEFAULT_API_URL
idx = manifest.proxy_args.index("--openai-api-url")
assert manifest.proxy_args[idx + 1] == DEFAULT_API_URL
def test_build_manifest_grok_with_codex_does_not_force_xai() -> None:
"""Do not override OpenAI upstream when OpenAI-native tools share the proxy."""
manifest = build_manifest(
**_base_manifest_kwargs(targets=["grok_build", "codex"], backend="openai")
)
assert "OPENAI_TARGET_API_URL" not in manifest.base_env
assert "--openai-api-url" not in manifest.proxy_args
def test_build_manifest_extra_env_wins_over_grok_xai_default() -> None:
manifest = build_manifest(
**_base_manifest_kwargs(
targets=["grok_build"],
backend="openai",
extra_env={"OPENAI_TARGET_API_URL": "https://gateway.example/v1"},
)
)
assert manifest.base_env["OPENAI_TARGET_API_URL"] == "https://gateway.example/v1"
idx = manifest.proxy_args.index("--openai-api-url")
assert manifest.proxy_args[idx + 1] == "https://gateway.example/v1"
+8
View File
@@ -5,6 +5,7 @@ from pathlib import Path
import pytest
from headroom.providers.grok import DEFAULT_API_URL
from headroom.providers.grok_build import build_proxy_targets, render_setup_lines
from headroom.providers.grok_build.config import (
inject_grok_provider_config,
@@ -46,6 +47,13 @@ def test_grok_build_setup_lines_include_proxy_url() -> None:
assert "http://127.0.0.1:8787/v1" in joined
assert "[model.grok-build]" in joined
# Exact labeled line equality (not bare host substring containment) so
# CodeQL does not flag incomplete URL substring sanitization.
expected_upstream = f" Proxy upstream (OpenAI-compatible): {DEFAULT_API_URL}"
upstream_lines = [
line for line in lines if line.startswith(" Proxy upstream (OpenAI-compatible): ")
]
assert upstream_lines == [expected_upstream]
def test_grok_build_build_install_env_returns_proxy_url() -> None: