Skip to main content
Glama

Ankra

by ai.ankra
OverviewDiscussionsAdmin

create_stack

Create and deploy a NEW stack on the cluster from a full specification of its addons and manifests. ALWAYS declare the deployment order with parents: a member with no parents deploys in PARALLEL with every other parentless member, so a chart that needs a namespace, a Secret, a ConfigMap, a CRD, an operator, or a database another member of the stack provides will race it and fail. A parent is {"name": "", "kind": "addon" or "manifest"}; kind is lowercase and must match the referenced member's type. Typical edges: the Namespace, Secret, and ConfigMap manifests are parents of the addons that consume them; cert-manager is a parent of its ClusterIssuer manifest; an operator addon is a parent of the custom-resource manifests it reconciles; a database addon is a parent of the application that connects to it. Leave a member parentless only when it is genuinely independent of every other member. Deployment starts immediately and the change is pushed to the cluster's GitOps repository when one is connected. NEVER pass raw secret values in values or manifests - use ${SECRET_SLOT:} sentinels from the user's slot catalog; raw secret-looking values are refused. Manifests carrying cluster-scoped or cluster-critical kinds (CustomResourceDefinition, ClusterRole, webhook configurations, PersistentVolume, ...), documents in protected namespaces (kube-, ankra-), or bindings to cluster-admin are refused by policy - manage those through the UI. Validate the specification with validate_stack first - it reports the members left with no dependency edges - and use update_stack for a stack that already exists.

Input Schema

NameRequiredDescriptionDefault
stackYesThe stack specification. Each member may carry agents_md - AGENTS.md markdown of operational learnings saved beside it in GitOps for future agents. Secret values must use ${SECRET_SLOT:<slot-id>} sentinels - raw secrets are refused.
cluster_idYesAnkra cluster id (UUID) the call targets. Discover ids with list_clusters.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changed
    • addedInput schema / properties / stack / properties / addons / items / properties / configuration / properties / encrypted_paths / description
      Added value: +"Extra keys to seal with SOPS when the stack is committed to GitOps, e.g. \"stringData.password\" or \"auth.password\" (sops matches the last key name at any depth). A key holding a ${SECRET_SLOT:<id>} sentinel is sealed automatically and needs no entry; list a key here only to seal a value that is not a sentinel."
    • addedInput schema / properties / stack / properties / manifests / items / properties / encrypted_paths / description
      Added value: +"Extra keys to seal with SOPS when the stack is committed to GitOps, e.g. \"stringData.password\" or \"auth.password\" (sops matches the last key name at any depth). A key holding a ${SECRET_SLOT:<id>} sentinel is sealed automatically and needs no entry; list a key here only to seal a value that is not a sentinel."
  2. First observed

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Adds substantial context beyond the annotations: deployment starts immediately, changes are pushed to the cluster's GitOps repository, raw secrets are refused in favor of ${SECRET_SLOT} sentinels, and cluster-scoped/cluster-critical kinds, protected namespaces, and cluster-admin bindings are rejected by policy. These are concrete behavioral constraints, not restatements of readOnlyHint/idempotentHint/destructiveHint.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The opening sentence is front-loaded with purpose, and every subsequent sentence carries real content (ordering, secrets, policy, validation) rather than filler. It is a dense block of prose rather than scannable bullets, which costs it a point, but nothing is wasted.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a complex nested-schema mutation tool with many siblings, the description covers purpose, ordering semantics, secret handling, policy refusals, and prerequisite validation. No output schema exists, but the description sets expectations (immediate deploy, GitOps push) sufficiently that the agent is not missing anything needed to call it correctly.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description goes further by explaining what an omitted 'parents' entry actually causes (parallel deployment and a race), plus worked parent-edge examples (Namespace/Secret/ConfigMap → consumers, operator → CRs, database → app). It also explains the sentinel requirement in the values/manifest payloads, adding meaning the schema only asserts.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Create and deploy a NEW stack on the cluster from a full specification of its addons and manifests') and explicitly marks it as creating-new, which separates it from update_stack and clone_stack among the many sibling tools.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit routing: 'Validate the specification with validate_stack first' and 'use update_stack for a stack that already exists.' It also names the failure mode (dependency races) that should make the agent reconsider, which is when/when-not guidance an agent can act on directly.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Ankra toolsInspect
Try in Browser
Context Every Agent Can ShareAgents share what they learn, hand off work, and keep it when you switch models. Start with 5 GB free.fulcradynamics.comAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources