AI agents invoke retry_pipeline to trigger actions in Gitlab. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
Retrying pipeline jobs re-executes build/test/deploy workloads in CI/CD. This is an Execute-category action as it triggers external operations (job runs) whose effects depend on the pipeline configuration. The blast radius is high because retried jobs may deploy code, modify infrastructure, or consume significant resources.
From the tool's definition "Retry failed jobs in a pipeline" — triggers re-execution of CI/CD pipeline jobs
Documented attack patterns abuse exactly the kind of access retry_pipeline gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Gitlab, and nothing reaches the server without passing your rules. This is the rule we recommend for retry_pipeline:
{
"version": "1",
"default": "deny",
"tools": {
"retry_pipeline": {
"limits": [
{
"counter": "retry_pipeline_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}retry_pipeline stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Retry failed jobs in a pipeline. It is categorised as a Execute tool in the Gitlab MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Gitlab MCP server in PolicyLayer and add a rule for retry_pipeline: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Gitlab. Nothing to install.
retry_pipeline is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the retry_pipeline rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for retry_pipeline. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
retry_pipeline is provided by the Gitlab MCP server (yoda-digital/mcp-gitlab-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 88 Gitlab tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
88 Gitlab tools catalogued and risk-classified — across an index of 42,500+ MCP servers.