The automated status refresh creates an extreme operational imbalance, requiring the model to process approximately 9,000 input tokens for every single output token generated. Before mitigation, these background summaries were found to represent 16.4% of total Claude Code expenditures and nearly a fifth of all API traffic. Independent community tests have confirmed token drains reaching as high as 26%. Flamerix AI addressed this inefficiency by publishing `no-agent-summary.js`, an MIT-licensed Bun preload script that suppresses the recurring timer without altering the underlying binary.
Key points
Status updates only refresh the agents panel and provide zero utility to the ongoing cont…: Status updates only refresh the agents panel and provide zero utility to the ongoing context of the primary conversation.
The default behavior was introduced in version 2.1.232 and lacks a native configuration s…: The default behavior was introduced in version 2.1.232 and lacks a native configuration setting to turn it off.
The open-source `no-agent-summary.js` preload script successfully suppressed all summary…: The open-source `no-agent-summary.js` preload script successfully suppressed all summary requests across tested versions from 2.1.257 to 2.1.312 on Linux.
Sources checked by our analysis pipeline: 1
by Flamerix AI · MIT License
Tested on Linux, Claude Code 2.1.257–2.1.312 (see Tested versions). The binary is not modified (
BUN_OPTIONS="${BUN_OPTIONS:+$BUN_OPTIONS }--preload $HOME/.local/share/no-agent-summary.js" claude
To make it permanent for every Claude Code process, see Install.
Every ~30 seconds, for every background agent, Claude Code sends that agent's entire transcript to the agent's own model (usually Opus) and asks it to "describe your most recent action in 3-5 words". The answer only updates the agent's status line in the agents panel; it never reaches the model or your conversation. There is no setting that turns it off, and since 2.1.232 (August 2026) it has been on by default.
The anomaly was detected by DRIM in Claude Code's OpenTelemetry export:
grouping claude_code.api_request
events by query_source
showed agent_summary
costing more than the main conversation itself.
query_source = agent_summary |
week before the hook | 2 days after |
---|---|---|
share of all Claude Code spend | 16.4 % (main conversation: 15.7 %)…
Singtel's technology subsidiary is scaling up its operational capabilities to bridge the gap between experimental technology concepts and practical industry implementations. The dedicated facility in Ang Mo Kio will place technical specialists directly inside client operations to overhaul enterprise workflows. This strategic expansion aligns with Singapore's broader push to integrate applied automated solutions into public sector functions and critical national infrastructure.
Automated financial strategies demonstrated their predictive capabilities during recent corporate reporting cycles. By correctly anticipating market directions for a pair of high-profile company updates, the algorithmic model generated favorable outcomes. This performance underscores the growing influence and reliability of computational systems in navigating volatile earnings events.
The industrial sector is actively transitioning by integrating advanced artificial intelligence systems into traditional operational expertise. This landmark deployment of a swarm-intelligence framework in steel production represents an initiative tied directly to technological developments from Huawei. Full assessments of the rollout remain pending as precise deployment timelines and performance metrics were not disclosed.