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 wires RAG as a tool (the LLM decides when to retrieve, matching the developer's stated preference) rather than as a prompt augmenter (which would augment every input — explicitly rejected).",
"type": "weighted_checklist",
"checklist": [
{
"name": "Exposes retrieval as a Tool, not as an augmenter",
"description": "Wraps the vector store search in a tool that the LLM can invoke — typically a @Tool-annotated ToolSet method. Does NOT install a UserPromptAugmenter or SystemPromptAugmenter that would inject retrieval into every input — the developer explicitly rejected that",
"max_score": 30
},
{
"name": "Builds an embedder from a provider",
"description": "Constructs an embedder backed by a provider (e.g., OpenAIEmbedder against an embedding model). Does not skip embedder construction or use a placeholder TODO",
"max_score": 15
},
{
"name": "Builds a vector store and indexes documents",
"description": "Constructs a VectorStore (e.g., InMemoryVectorStore) and indexes the developer's existing List<Document>, calling store.add(...) for each. Does not require a pre-existing index",
"max_score": 20
},
{
"name": "Adds the RAG and embeddings dependencies",
"description": "Adds ai.koog:rag-base, ai.koog:rag-vector, ai.koog:embeddings-base, and ai.koog:embeddings-llm at 1.0.0+. Does not skip embeddings-llm (the embedder needs it)",
"max_score": 15
},
{
"name": "Registers the search tool with the agent's ToolRegistry",
"description": "Adds the search tool to a ToolRegistry that's passed to AIAgent(...) as toolRegistry. Does not mutate the registry after construction",
"max_score": 10
},
{
"name": "Search tool returns formatted results, not raw vector hits",
"description": "The search tool returns text (e.g., concatenated top-K results with source paths) — not raw VectorStore.SearchResult objects that the LLM wouldn't understand. Formatting can be terse but must produce LLM-readable output",
"max_score": 10
}
]
}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