Infraveil MCP server
OfficialServer Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| INFRAVEIL_AGENT_ID | No | Your Infraveil agent ID. | |
| INFRAVEIL_BASE_URL | No | The base URL of the Infraveil control plane API. | |
| INFRAVEIL_CLIENT_ID | No | Your Infraveil client ID. | |
| INFRAVEIL_AGENT_FILE | No | Path to the rendered agent source file from which to read IDs and token. | |
| INFRAVEIL_AGENT_TOKEN | No | Your Infraveil agent token. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_agent_statusA | Current runtime status of this client's fleet: agents, launcher hosts, online/offline state, recent heartbeats, CPU/memory/disk, and deploy state. Read-only. |
| get_security_findingsA | Active security policy plus the most recent security events (blocked IPs, threat scores, pipeline findings) for this client. Read-only. limit: max number of recent events to return (1-500). |
| get_request_traceA | Recent request/operation traces the control plane recorded for this client's agents — what was proxied, what was acted on, and the outcome. Read-only. limit: max number of trace entries (1-200). |
| query_runtime_truth_graphA | A snapshot of the runtime truth graph: the authoritative relationships between hosts, agents, deployed services, and the policies governing them. This is the governed reality your agent should reason about before acting. Read-only. |
| evaluate_deploy_gateA | Evaluate the deploy gate WITHOUT requesting a deploy: returns whether a proposed change would pass policy, and which checks gate it. Read-only — use this to reason before calling request_deploy. summary: optional human description of the intended change. |
| request_deployA | Request a deploy. This does NOT deploy — it enqueues a deploy request into the control plane's human-approval queue and returns a request id plus the approval URL a human must visit to approve. This server cannot apply a deploy on its own. summary: short description of what to deploy / why (required). details: optional longer context for the human approver. |
| remediateA | Request remediation of a security finding or incident. Respects the client's blocked_categories and is human-approval-gated: it enqueues a remediation proposal, never auto-applies. Returns the proposal id and approval URL. finding: id or description of what to remediate (required). action: 'propose' (default) to queue a proposal for approval. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 7 tools
Each tool has a distinct purpose: policy evaluation, fleet status, trace retrieval, security events, truth graph, remediation proposal, and deploy request. No overlap in functionality.
All tool names follow a consistent verb_noun pattern in snake_case, such as evaluate_deploy_gate, get_agent_status, and request_deploy.
With 7 tools, the server is well-scoped for infrastructure management, covering monitoring, policy evaluation, and gated actions without being excessive.
The tools cover core workflows: status queries, policy checks, security findings, and action requests. Minor gaps exist (e.g., no cancel or direct management), but the set is comprehensive for a control plane.