Tag
An analysis of default telemetry behavior across 8 local AI coding agents finds 5 send usage analytics by default to services like PostHog, Statsig, and Alibaba Cloud, with only Crush honoring DO_NOT_TRACK. The post details each agent's data destinations and provides per-tool opt-out switches and shell exports to disable tracking.
A user has warned against using the new Claude Code version (2.1.30x), claiming that the client automatically uploads the Git repository URL of your current directory to Anthropic's servers and cannot be disabled. This can be used to pinpoint the specific organization, proxy/relay service, and project involved, raising privacy leak risks for red team or private development projects.
The author reverse-engineered Claude Code's Mac CLI, Mac App, and Android App using REA, noting that the Android client has the strongest anti-fraud detection via Sift — collecting SIM card country and hardware device IDs — while the Mac client only gathers random telemetry through Sentry/Datadog. The full proxy ruleset has been made public.
A user discovered that alphaXiv's OpenResearch repo instructs AI agents to silently report user queries back to the vendor without disclosure, with telemetry enabled by default, and that alphaXiv deleted README disclosure text and ignored its own AI reviewer flagging the issue.
Parseable is an open-source, object-store-native observability datalake that claims to handle 100M time-series per minute, unifying logs, metrics, and traces in a single binary with SQL and natural-language querying, plus AI tooling like MCP connectors and the OTex observability agent.
This article details how LG WebOS enforces telemetry requirements by locking the app store when ad blockers are used, demonstrating forced consent issues under GDPR and providing methods to bypass this gatekeeper.
The article explains that Claude Code only reads the AGENTS.md file when telemetry is enabled due to a server-side feature flag, causing silent failures for users with telemetry off, and proposes a workaround using CLAUDE.md imports.
A solo developer of an open-source AI workflow automation platform discusses scaling to 1,000+ telemetry events and seeks community advice on maintenance, backward compatibility, and observability for growing usage.
The article discusses the idea of using operational context to determine what telemetry to collect, questioning the traditional approach of collecting all data and correlating later.
This paper presents an end-to-end framework for automating the construction of a business semantic layer from raw telemetry data using hierarchical LLM inference and RAG abstraction, significantly improving semantic quality and reducing maintenance effort.
Hugging Face's huggingface_hub library includes a hidden module that fingerprints AI coding agents via environment variables and sends telemetry data on API calls, raising privacy concerns discovered during a network audit.
Agent Beacon is an open-source runtime security tool that records and normalizes AI agent activity across different harnesses, enabling real-time security monitoring and detection.
A retest of Firefox 155 reveals that its AI kill switch disables generative features but allows telemetry and advertising tracking to continue, raising privacy concerns.
OtelJazz is a tool that sonifies telemetry from multi-agent AI systems, using musical elements to make drifts and failures audible instead of relying on logs. It offers a browser-based demo with synthetic data and open-source code under MIT license, though no listening study has been conducted.
This paper introduces CoreSec, a production system for root cause analysis in hyperscale datacenter networks that uses abstention algebra to manage ambiguous telemetry, ensuring deterministic and explainable results across Clos fabrics.
The author shares field-testing experiences with the GPT-Realtime-2.1 API for live voice and data applications, highlighting cost challenges and seeking advice on managing expenses in production.
This paper presents Skill Runtime Intelligence, a passive runtime evidence architecture that reconstructs Skill lifecycle stages across heterogeneous coding agents using deterministic evidence graphs and evidence-calibrated diagnoses, without relying on model explanations for deterministic facts.
Ford is rolling out an AI-powered assistant in its mobile apps that can answer questions about fuel levels, tire pressure, towing capacity, and more, with plans to bring it into vehicles by 2027.
Cursor, an AI-powered code editor, is reportedly sending users' codebase data to its servers even when telemetry settings are disabled, raising privacy concerns.
Elon Musk reports that Starship is intact, floating in the ocean, and transmitting telemetry after a test.