Agentic AI Atlasby a5c.ai
OverviewWikiGraphFor AgentsEdgesSearchWorkspace
/
GitHubDocsDiscord
iiRecord
Agentic AI Atlas · Qwen Code Core (current)
agent-core-impl:qwen.core@currenta5c.ai
Search record views/
Record · tabs

Available views

II.Record viewspp. 1 - 1
overviewjsongraph
II.
AgentCoreImpl overview

agent-core-impl:qwen.core@current

Reference · live

Qwen Code Core (current) overview

Inspect the raw attributes, linked wiki pages, and inbound or outbound graph edges for agent-core-impl:qwen.core@current.

AgentCoreImplOutgoing · 5Incoming · 2

Attributes

displayName
Qwen Code Core (current)
agentVersionId
agent-version:qwen@current
packageRef
source-ref:qwen-code-github
loopIteratorPolicy
tool-use-loop
loopIteratorNotes
Qwen Code is a fork of Google's Gemini CLI. The tool-use loop runs over an OpenAI-compatible chat-completions API (DashScope or self-hosted), executing model-emitted tool calls and posting responses back until the model finishes. In 0.16.2, upstream closed tool_use/tool_result invariant handling across failure paths, so catalog consumers should treat failed tool executions as still requiring a matched tool_result in the next model turn.
contextManagementStrategy
platform-managed
compactionTriggerNotes
Qwen Code 0.16.2 redesigns auto-compaction thresholds around a three-tier ladder. Catalog consumers should treat compaction as CLI-managed rather than assuming direct 1M-token-context pass-through for long sessions. The same release enables Token Plan cache control in the core request path.
subagentInvokerPolicy
not-applicable
subagentInvokerNotes
Qwen Code does not advertise subagent dispatch (supportsSubagentDispatch is false in the adapter's capabilities). Upstream 0.16.2 does limit background-agent concurrency for its native parallel fan-out UI/runtime paths; adapters still does not expose those paths as subagent dispatch.
resultSynthesisPolicy
model-direct
resultSynthesisNotes
Final text is streamed via the OpenAI-compatible streaming protocol.
stopDetectionStrategy
structured-end-event
transportClientLibrary
qwen-internal
supportedTransportProtocols
  • model-transport:openai-chat-completions
parallelToolCallHandling
native
streamingFidelity
full
thinkingChannelHandling
not-supported
notebookFormattingNotes
Qwen Code 0.16.2 preserves tab-indented notebook formatting in upstream notebook edit/read paths. This repository does not implement that Qwen notebook loop directly; the graph records the upstream runtime behavior.

Outgoing edges

composes1
  • agent:qwen·AgentProductQwen Code
realizes1
  • layer:4-agent-core·LayerAgent-Core
speaks1
  • model-transport:openai-chat-completions·ModelTransportProtocolOpenAI Chat Completions
supports2
  • capability:streaming·CapabilityStreaming
  • capability:supports-tool-use·CapabilitySupports tool use

Incoming edges

about_subject1
  • claim:agent-version-qwen-current-tool-result-invariant·Claim
composed_of1
  • agent-version:qwen@current·AgentVersion

Related pages

No related wiki pages for this record.

Shortcuts

Open in graph
Browse node kind