Add Claude Code remote environment setup automation #3421
Conversation
This adds automatic installation of system dependencies when running in Claude Code remote (web) environments: - protobuf-compiler for Rust protobuf compilation - librdkafka-dev and related libs for Kafka JavaScript module - Python 3.12 for moose-cli pip package Note: Docker is not available in Claude Code remote environments, so E2E tests requiring Docker must run in CI or locally. https://claude.ai/code/session_0135cswZZnwLDYtR9f3FkBNa
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
WalkthroughThis PR introduces automated setup for Claude Code remote environments by adding a SessionStart hook configuration, documentation of pre-installed dependencies, and a comprehensive Bash script that installs Docker, system libraries, protobuf-compiler, Kafka dependencies, and Python 3.12+ with configuration. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
- Add VERSION CONSTANTS section at top of script for easy updates - Add Docker and Docker Compose installation with fallback handling - Update AGENTS.md to reflect Docker installation attempt https://claude.ai/code/session_0135cswZZnwLDYtR9f3FkBNa
- Add version constants for protobuf-compiler, librdkafka-dev, libcurl4-openssl-dev, and libsasl2-dev - Use versioned package installation with fallback to latest - Show actual installed versions in summary output https://claude.ai/code/session_0135cswZZnwLDYtR9f3FkBNa
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Fix all issues with AI agents
In `@AGENTS.md`:
- Line 95: Add a single trailing newline at end of AGENTS.md so the file ends
with one newline (fixes Markdownlint MD047); open AGENTS.md and ensure that
after the final line "Run full E2E tests (`cd apps/framework-cli-e2e && pnpm
test`) in environments with Docker support" there is exactly one newline
character at EOF.
- Around line 82-95: Add a new short section titled "Agent architecture / design
patterns" to AGENTS.md (place it after the "Claude Code Remote Environment"
section) that either summarizes the canonical agent architecture or links to the
authoritative design doc; include one-paragraph descriptions of the recommended
patterns to follow: message-passing vs shared-memory, controller/worker
separation, stateless vs stateful agents and state persistence, retry/backoff
and error-handling conventions, and observability/metrics/logging practices,
plus a link to the primary design document or RFC for deeper details.
In `@scripts/claude-remote-setup.sh`:
- Around line 30-34: The apt-get guard should also check for sufficient
privileges before attempting package installation; update the logic around the
existing apt-get check in the script (scripts/claude-remote-setup.sh) to first
verify the script is running as root (UID 0) and if not, either attempt to use
sudo (if sudo is available) for apt-get commands or skip system package
installation and exit 0 gracefully—ensure this privilege check runs before the
current "if ! command -v apt-get" block and that any error paths do not trigger
set -e to abort the entire setup.
📜 Review details
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro
Disabled knowledge base sources:
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (3)
.claude/settings.jsonAGENTS.mdscripts/claude-remote-setup.sh
🧰 Additional context used
📓 Path-based instructions (1)
AGENTS.md
📄 CodeRabbit inference engine (CLAUDE.md)
Document agent architecture and design patterns in AGENTS.md
Files:
AGENTS.md
🧠 Learnings (4)
📚 Learning: 2025-12-16T23:08:46.806Z
Learnt from: CR
Repo: 514-labs/moosestack PR: 0
File: packages/protobuf/CLAUDE.md:0-0
Timestamp: 2025-12-16T23:08:46.806Z
Learning: Applies to packages/protobuf/**/AGENTS.md : Review AGENTS.md for coding standards and guidelines
Applied to files:
AGENTS.md
📚 Learning: 2025-12-16T23:08:39.003Z
Learnt from: CR
Repo: 514-labs/moosestack PR: 0
File: apps/framework-cli/CLAUDE.md:0-0
Timestamp: 2025-12-16T23:08:39.003Z
Learning: Document agent implementations in AGENTS.md file
Applied to files:
AGENTS.md
📚 Learning: 2025-12-16T23:08:34.942Z
Learnt from: CR
Repo: 514-labs/moosestack PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-16T23:08:34.942Z
Learning: Applies to AGENTS.md : Document agent architecture and design patterns in AGENTS.md
Applied to files:
AGENTS.md
📚 Learning: 2025-12-16T23:08:31.521Z
Learnt from: CR
Repo: 514-labs/moosestack PR: 0
File: .github/copilot-instructions.md:0-0
Timestamp: 2025-12-16T23:08:31.521Z
Learning: Ensure all tests pass: `cargo test` for Rust components and appropriate test commands for TS/JS
Applied to files:
AGENTS.md
🪛 markdownlint-cli2 (0.20.0)
AGENTS.md
[warning] 95-95: Files should end with a single newline character
(MD047, single-trailing-newline)
🪛 Shellcheck (0.11.0)
scripts/claude-remote-setup.sh
[info] 61-61: Not following: /etc/os-release was not specified as input (see shellcheck -x).
(SC1091)
🔇 Additional comments (1)
.claude/settings.json (1)
5-10: The code is correct. Claude Code automatically injectsCLAUDE_PROJECT_DIRand expands environment variables in hook commands, which run as regular shell commands. The escaped quotes are the correct JSON syntax for embedding a shell command with variable expansion. No changes needed.Likely an incorrect or invalid review comment.
✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.
| ## Claude Code Remote Environment | ||
|
|
||
| When running in Claude Code remote (web) environments, system dependencies are automatically installed via the SessionStart hook in `.claude/settings.json`. The setup script (`scripts/claude-remote-setup.sh`) installs: | ||
|
|
||
| - **Docker & Docker Compose**: Required for E2E tests (ClickHouse, Kafka, Temporal containers) | ||
| - **protobuf-compiler**: Required for Rust build (prost/tonic protobuf compilation) | ||
| - **librdkafka-dev**: Required for `@514labs/kafka-javascript` native module | ||
| - **libcurl4-openssl-dev, libsasl2-dev**: Additional Kafka dependencies | ||
| - **Python 3.12+**: Required for `moose-cli` pip package in Python E2E tests | ||
|
|
||
| **Limitations in Claude Code Remote:** | ||
| - **Docker may not be available**: Docker requires privileged container access which may not be supported in all environments | ||
| - Run unit tests (`cargo test`, `pnpm test` in ts-moose-lib) and linting in remote environments | ||
| - Run full E2E tests (`cd apps/framework-cli-e2e && pnpm test`) in environments with Docker support No newline at end of file |
There was a problem hiding this comment.
Add agent architecture/design patterns section
This update doesn’t document agent architecture/design patterns as required; please add a short section or link to the authoritative doc.
As per coding guidelines: Document agent architecture and design patterns in AGENTS.md.
🧰 Tools
🪛 markdownlint-cli2 (0.20.0)
[warning] 95-95: Files should end with a single newline character
(MD047, single-trailing-newline)
🤖 Prompt for AI Agents
In `@AGENTS.md` around lines 82 - 95, Add a new short section titled "Agent
architecture / design patterns" to AGENTS.md (place it after the "Claude Code
Remote Environment" section) that either summarizes the canonical agent
architecture or links to the authoritative design doc; include one-paragraph
descriptions of the recommended patterns to follow: message-passing vs
shared-memory, controller/worker separation, stateless vs stateful agents and
state persistence, retry/backoff and error-handling conventions, and
observability/metrics/logging practices, plus a link to the primary design
document or RFC for deeper details.
| **Limitations in Claude Code Remote:** | ||
| - **Docker may not be available**: Docker requires privileged container access which may not be supported in all environments | ||
| - Run unit tests (`cargo test`, `pnpm test` in ts-moose-lib) and linting in remote environments | ||
| - Run full E2E tests (`cd apps/framework-cli-e2e && pnpm test`) in environments with Docker support No newline at end of file |
There was a problem hiding this comment.
Add trailing newline at EOF
Markdownlint MD047 expects a single trailing newline after Line 95.
Proposed fix
- - Run full E2E tests (`cd apps/framework-cli-e2e && pnpm test`) in environments with Docker support
+ - Run full E2E tests (`cd apps/framework-cli-e2e && pnpm test`) in environments with Docker support
+📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| - Run full E2E tests (`cd apps/framework-cli-e2e && pnpm test`) in environments with Docker support | |
| - Run full E2E tests (`cd apps/framework-cli-e2e && pnpm test`) in environments with Docker support | |
🧰 Tools
🪛 markdownlint-cli2 (0.20.0)
[warning] 95-95: Files should end with a single newline character
(MD047, single-trailing-newline)
🤖 Prompt for AI Agents
In `@AGENTS.md` at line 95, Add a single trailing newline at end of AGENTS.md so
the file ends with one newline (fixes Markdownlint MD047); open AGENTS.md and
ensure that after the final line "Run full E2E tests (`cd apps/framework-cli-e2e
&& pnpm test`) in environments with Docker support" there is exactly one newline
character at EOF.
| # Check if we have apt-get available | ||
| if ! command -v apt-get &> /dev/null; then | ||
| echo "apt-get not found, skipping system package installation" | ||
| exit 0 | ||
| fi |
There was a problem hiding this comment.
Guard against non-root execution before apt-get
With set -e, running the hook without root will abort setup; add a privilege check (or sudo path) and skip gracefully when not root.
Proposed fix
if ! command -v apt-get &> /dev/null; then
echo "apt-get not found, skipping system package installation"
exit 0
fi
+if [ "$(id -u)" -ne 0 ]; then
+ echo "Insufficient privileges to install packages; skipping setup."
+ exit 0
+fi📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| # Check if we have apt-get available | |
| if ! command -v apt-get &> /dev/null; then | |
| echo "apt-get not found, skipping system package installation" | |
| exit 0 | |
| fi | |
| # Check if we have apt-get available | |
| if ! command -v apt-get &> /dev/null; then | |
| echo "apt-get not found, skipping system package installation" | |
| exit 0 | |
| fi | |
| if [ "$(id -u)" -ne 0 ]; then | |
| echo "Insufficient privileges to install packages; skipping setup." | |
| exit 0 | |
| fi |
🤖 Prompt for AI Agents
In `@scripts/claude-remote-setup.sh` around lines 30 - 34, The apt-get guard
should also check for sufficient privileges before attempting package
installation; update the logic around the existing apt-get check in the script
(scripts/claude-remote-setup.sh) to first verify the script is running as root
(UID 0) and if not, either attempt to use sudo (if sudo is available) for
apt-get commands or skip system package installation and exit 0
gracefully—ensure this privilege check runs before the current "if ! command -v
apt-get" block and that any error paths do not trigger set -e to abort the
entire setup.
The deadsnakes PPA may be blocked by network proxies in some environments. Update the script to catch this failure and continue with a warning instead of failing the entire setup. https://claude.ai/code/session_0135cswZZnwLDYtR9f3FkBNa
Summary
This PR adds automatic setup for Claude Code remote (web) environments by introducing a SessionStart hook that installs required system dependencies and documents environment limitations.
Key Changes
.claude/settings.json: Configures a SessionStart hook to automatically run the setup script when Claude Code remote sessions startscripts/claude-remote-setup.sh: Bash script that installs critical system dependencies:protobuf-compilerfor Rust builds (prost/tonic)librdkafka-dev,libcurl4-openssl-dev,libsasl2-devfor Kafka native modulespython3.12+for moose-cli pip package and E2E testsAGENTS.md: Added documentation section explaining:Implementation Details
CLAUDE_CODE_REMOTEenv var)apt-getand skips if not availableThis enables developers to work effectively in Claude Code remote environments while being aware of constraints around containerized services.
https://claude.ai/code/session_0135cswZZnwLDYtR9f3FkBNa
Note
Medium Risk
Adds an auto-run SessionStart hook that executes a privileged
apt-get/repo-setup script; risk is mainly around unexpected environment side effects and network/package availability, not product logic.Overview
Automates Claude Code remote setup. Adds
.claude/settings.jsonto runscripts/claude-remote-setup.shon SessionStart (guarded byCLAUDE_CODE_REMOTE=true).The setup script installs build/test prerequisites via
apt-get(Docker/Compose,protobuf-compiler, Kafka native deps, and optionally Python 3.12 via deadsnakes), with best-effort fallbacks and a summary output.Updates
AGENTS.mdto document what gets installed in remote sessions and the Docker availability limitation, plus guidance on which tests to run remotely vs. in Docker-capable environments.Written by Cursor Bugbot for commit f57cea2. This will update automatically on new commits. Configure here.