External agent configuration examples
- Tier: Premium, Ultimate
- Add-on: GitLab Duo Core, Pro, or Enterprise, GitLab Duo with Amazon Q
- Offering: GitLab.com, GitLab Self-Managed, GitLab Dedicated
- Status: Experiment
Version history
- Introduced in GitLab 18.3 with a flag named
ai_flow_triggers. Enabled by default.
Use the following examples to create your external agent configuration. These examples contain the following variables:
-
AI_FLOW_CONTEXT: the JSON-serialized parent object, including:- In merge requests, the diff and comments (up to a limit)
- In issues or epics, the comments (up to a limit)
-
$AI_FLOW_EVENT: the type of trigger event (for example,mention) -
$AI_FLOW_INPUT: the prompt the user enters as a comment in the merge request, issue, or epic
Integrated with GitLab
The following agents are integrated with GitLab and available on GitLab.com.
Anthropic Claude
injectGatewayToken: true
image: node:22-slim
commands:
- echo "Installing claude"
- npm install --global @anthropic-ai/claude-code
- echo "Installing glab"
- export GITLAB_TOKEN=$GITLAB_TOKEN_CLAUDE
- apt-get update --quiet && apt-get install --yes curl wget gpg git && rm --recursive --force /var/lib/apt/lists/*
- curl --silent --show-error --location "https://raw.githubusercontent.com/upciti/wakemeops/main/assets/install_repository" | bash
- apt-get install --yes glab
- echo "Configuring git"
- git config --global user.email "claudecode@gitlab.com"
- git config --global user.name "Claude Code"
- echo "Configuring claude"
- export ANTHROPIC_AUTH_TOKEN=$AI_FLOW_AI_GATEWAY_TOKEN
- export ANTHROPIC_CUSTOM_HEADERS=$AI_FLOW_AI_GATEWAY_HEADERS
- export ANTHROPIC_BASE_URL="https://cloud.gitlab.com/ai/v1/proxy/anthropic"
- echo "Running claude"
- |
claude --debug --allowedTools="Bash(glab:*),Bash(git:*)" --permission-mode acceptEdits --verbose --output-format stream-json -p "
You are an AI assistant helping with GitLab operations.
Context: $AI_FLOW_CONTEXT
Task: $AI_FLOW_INPUT
Event: $AI_FLOW_EVENT
Please execute the requested task using the available GitLab tools.
Be thorough in your analysis and provide clear explanations.
<important>
Use the glab CLI to access data from GitLab. The glab CLI has already been authenticated. You can run the corresponding commands.
When you complete your work create a new Git branch, if you aren't already working on a feature branch, with the format of 'feature/<short description of feature>' and check in/push code.
When you check in and push code, you will need to use the access token stored in GITLAB_TOKEN and the user ClaudeCode.
Lastly, after pushing the code, if a merge request doesn't already exist, create a new merge request for the branch and link it to the issue using:
`glab mr create --title "<title>" --description "<desc>" --source-branch <branch> --target-branch <branch>`
If you are asked to summarize a merge request or issue, or asked to provide more information, then please post back a note to the merge request / issue so that the user can see it.
</important>
"
variables:
- GITLAB_TOKEN_CLAUDE
- GITLAB_HOST
OpenAI Codex
image: node:22-slim
injectGatewayToken: true
commands:
- echo "Installing codex"
- npm install --global @openai/codex
- echo "Installing glab"
- export OPENAI_API_KEY=$AI_FLOW_AI_GATEWAY_TOKEN
- export GITLAB_TOKEN=$GITLAB_TOKEN_CODEX
- apt-get update --quiet && apt-get install --yes curl wget gpg git && rm --recursive --force /var/lib/apt/lists/*
- curl --silent --show-error --location "https://raw.githubusercontent.com/upciti/wakemeops/main/assets/install_repository" | bash
- apt-get install --yes glab
- echo "Configuring git"
- git config --global user.email "codex@gitlab.com"
- git config --global user.name "OpenAI Codex"
- echo "Running Codex"
- |
# Parse AI_FLOW_AI_GATEWAY_HEADERS (newline-separated "Key: Value" pairs)
header_str="{"
first=true
while IFS= read -r line; do
# skip empty lines
[ -z "$line" ] && continue
key="${line%%:*}"
value="${line#*: }"
if [ "$first" = true ]; then
first=false
else
header_str+=", "
fi
header_str+="\"$key\" = \"$value\""
done <<< "$AI_FLOW_AI_GATEWAY_HEADERS"
header_str+="}"
codex exec \
--config 'model_provider="gitlab"' \
--config 'model_providers.gitlab.name="GitLab Managed Codex"' \
--config 'model_providers.gitlab.base_url="https://cloud.gitlab.com/ai/v1/proxy/openai/v1"' \
--config 'model_providers.gitlab.env_key="OPENAI_API_KEY"' \
--config 'model_providers.gitlab.wire_api="responses"' \
--config "model_providers.gitlab.http_headers=${header_str}" \
--config shell_environment_policy.ignore_default_excludes=true \
--dangerously-bypass-approvals-and-sandbox "
You are an AI assistant helping with GitLab operations.
Context: $AI_FLOW_CONTEXT
Task: $AI_FLOW_INPUT
Event: $AI_FLOW_EVENT
Please execute the requested task using the available GitLab tools.
Be thorough in your analysis and provide clear explanations.
<important>
Use the glab CLI to access data from GitLab. The glab CLI has already been authenticated. You can run the corresponding commands.
When you complete your work create a new Git branch, if you aren't already working on a feature branch, with the format of 'feature/<short description of feature>' and check in/push code.
When you check in and push code, you will need to use the access token stored in GITLAB_TOKEN and the user Codex.
Lastly, after pushing the code, if a merge request doesn't already exist, create a new merge request for the branch and link it to the issue using:
glab mr create --title \"<title>\" --description \"<desc>\" --source-branch \"<branch>\" --target-branch \"<branch>\"
If you are asked to summarize a merge request or issue, or asked to provide more information then please post back a note to the merge request / issue so that the user can see it.
</important>
"
variables:
- GITLAB_TOKEN_CODEX
- GITLAB_HOST