Skip to content

Latest commit

 

History

History
139 lines (101 loc) · 5.07 KB

TOOLS.md

File metadata and controls

139 lines (101 loc) · 5.07 KB

MCP Server Tools Catalog

A concise, discoverable catalog of all MCP tools exposed by this server. Use the list_tools meta tool for an always up-to-date runtime view.

🔍 Quick Discovery

From a Copilot / MCP chat, run:

Tool: list_tools
Args: {}

Filter by substring:

Tool: list_tools
Args: { "filter": "deploy" }

🧰 Tool Categories

Category Tools
Cluster & Inventory get_cluster_info, list_namespaces, list_deployments
Deployment Lifecycle get_deployment_status, scale_deployment, deploy_version, rollback_deployment
Debug & Diagnostics get_pod_logs
Feature Management toggle_feature_flag
Policy Evaluation evaluate_deployment_policies, list_policy_rules, auto_fix_policy_violations
Compliance & Reporting generate_compliance_report, preview_policy_impact
Policy Generation & Tuning generate_policy_configuration, validate_policy_configuration, suggest_policy_customizations
Meta / Discovery list_tools, natural_language_command

📦 Detailed Reference

Cluster & Inventory

get_cluster_info

Returns cluster meta (name, server URL, version, counts). Use before other actions to verify connectivity.

list_namespaces

Lists namespaces with deployment counts and status.

list_deployments

Lists deployments optionally scoped by namespace with health summary.

Deployment Lifecycle

get_deployment_status

Shows replica readiness, conditions, rollout state, optional events.

scale_deployment

Horizontally scale a deployment with optional wait for readiness.

deploy_version

Update deployment image with strategy (RollingUpdate/Recreate) and progress stats.

rollback_deployment

Rollback to previous or specific revision; reports images and timings.

Debug & Diagnostics

get_pod_logs

Fetch recent pod logs for a deployment; supports container selection, previous instance and tail length.

Feature Management

toggle_feature_flag

Enable/disable a feature flag via env var or ConfigMap injection.

Policy Evaluation

evaluate_deployment_policies

Evaluate a single deployment; returns violations + warnings + summary.

list_policy_rules

List all rules (or by category) with severity and auto-fix info.

auto_fix_policy_violations

Apply safe remediation steps for fixable violations (dry-run supported via future enhancement).

Compliance & Reporting

generate_compliance_report

Aggregate compliance for a namespace or whole cluster.

preview_policy_impact

What-if comparison between current and alternate config file.

Policy Generation & Tuning

generate_policy_configuration

Produce baseline config JSON from high-level org inputs.

validate_policy_configuration

Check a config file structure and rule references.

suggest_policy_customizations

Analyze repeated violations to recommend overrides (strict/advisory, enable/disable).

Meta

list_tools

Enumerate all registered tools, their parameters, and example invocation objects.

natural_language_command

Parses a free‑form English instruction and returns the inferred tool + structured parameters (does NOT automatically execute the underlying action to avoid unintended changes). Use when you want to explore or prototype without memorizing tool names.

Example queries:

"Scale deployment cart-api in namespace shop to 5 replicas"
"Show me the status of deployment payments in ns prod"
"Deploy image ghcr.io/acme/api:v2.3.1 to deployment api-gateway in namespace edge"
"Generate a compliance report for namespace finance"

If required parameters are missing you’ll get a prompt listing them. Ambiguous or unknown requests return suggested alternative tools.

🗺 Usage Patterns

Goal Recommended Sequence
Onboard / explore list_tools → get_cluster_info → list_namespaces → list_deployments
Scale app get_deployment_status → scale_deployment → get_deployment_status
Deploy new version deploy_version → get_deployment_status → rollback_deployment (if needed)
Investigate issue get_deployment_status → get_pod_logs → evaluate_deployment_policies
Improve compliance generate_compliance_report → suggest_policy_customizations → preview_policy_impact
Create new policy config generate_policy_configuration → validate_policy_configuration → preview_policy_impact
Auto remediate evaluate_deployment_policies → auto_fix_policy_violations → re-run evaluate

💡 Best Practices

  • Use list_tools periodically—future versions may add tools without code browsing.
  • Always start with get_cluster_info to surface connectivity issues early.
  • Run generate_compliance_report before large-scale changes to establish a baseline.
  • Use preview_policy_impact before swapping configs in production.
  • Prefer advisory enforcement first; escalate to strict once noise is reduced.

🔄 Keeping Docs Current

This file is static; the canonical runtime source of truth is the list_tools meta tool. Update this file when adding or renaming tools for external consumers.