Koog 1.0 idioms, gotchas, and scaffolding skills for Kotlin agents on the JVM
88
88%
Does it follow best practices?
Impact
88%
1.95xAverage score across 43 eval scenarios
Passed
No known issues
{
"context": "Tests whether the agent reaches for the planner's PlannerNode tree (the right depth — parallel + sequential composition modeled explicitly) AND uses storage keys to track in-flight subtasks for the debugging requirement — rather than building this as a graph strategy with hand-wired conditional edges.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Uses PlannerNode.parallel inside PlannerNode.sequential",
"description": "Composes the plan via PlannerNode.sequential(...) wrapping PlannerNode.parallel(...) — explicit tree composition. Does NOT model the angles as conditional edges in a strategy { ... } graph (which serializes them) and does NOT use PlannerNode.sequential alone (would serialize the angles)",
"max_score": 30
},
{
"name": "Builds the agent with PlannerAIAgent / Planners.*",
"description": "Constructs the agent through the top-level AIAgent(...) factory with a strategy from Planners.llmBased or AIAgentPlannerStrategy(..., planner). Does NOT use GraphAIAgent (which would mean hand-rolling planner behavior in a graph)",
"max_score": 20
},
{
"name": "Tracks current subtask via storage key",
"description": "Declares a typed storage key (createStorageKey<PlannerNode.Builder.Reference>(\"...\")) and stores the current subtask reference on entry. Without this the debugging requirement (\"know which subtask is executing\") is unmet",
"max_score": 20
},
{
"name": "Adds the planner module dependency",
"description": "Adds ai.koog:agents-planner at 1.0.0+ to build.gradle.kts. The umbrella does not include it",
"max_score": 15
},
{
"name": "Sets maxIterations well above the default",
"description": "Raises maxIterations significantly above the factory default of 50 (e.g., 200+). Each subtask plus the planning step costs multiple iterations; a tree of 4 angles plus composition needs much more than 50",
"max_score": 10
},
{
"name": "Does not over-engineer with graph subgraphs",
"description": "Does not author a strategy { ... } block with subgraphWithTask on top of the planner — the planner already handles the tree internally; adding subgraphs on top is double-modeling",
"max_score": 5
}
]
}evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
scenario-11
scenario-12
scenario-13
scenario-14
scenario-15
scenario-16
scenario-17
scenario-18
scenario-19
scenario-20
scenario-21
scenario-22
scenario-23
scenario-24
scenario-25
scenario-26
scenario-27
scenario-28
scenario-29
scenario-30
scenario-31
scenario-32
scenario-33
scenario-34
scenario-35
scenario-36
scenario-37
scenario-38
scenario-39
scenario-40
scenario-41
scenario-42
scenario-43
skills
add-observability
add-persistence
add-rag
add-structured-output
add-token-budgeting
add-tool
cache-llm-calls
define-prompt
domain-model-subtask-pipeline
references
enable-prompt-caching
handle-agent-events
manage-state
migrate-from-0-x
model-planner-subtasks
persist-chat-history
query-sql-from-agent
scaffold-agent
snapshot-and-restore
test-koog-agents
trace-agent-internals
use-attachments
use-functional-agent
use-llm-node-variants
use-planner
wire-a2a
wire-acp-server
wire-ktor-server
wire-mcp-server
wire-spring-boot