Agents & Frameworks
Cloudflare Precursor: Behavioral Detection for AI Agents
Cloudflare Precursor reportedly shifts AI agent detection from spoofable headers to continuous behavioral signals. This article verifies the claims and outlines operator.
The Groundy archive · Page 12 of 38
All articles
Browse Groundy's complete archive of 903 articles on AI, developer tools and infrastructure. Page 12 of 38.
The archive
265–288 of 903 articles · Newest first
- Developer Tools
Provenance as a CI Gate: Attributing Agent-Authorship in Code
Git credits the merger, not the model. F(AI)2R proposes a PROV-O provenance graph gated by CI to record agent authorship and human verification, preventing blame collapse in.
Developer ToolsOHTTP CLI: Stateless Privacy for Agents vs VPN and Tor
Cloudflare's OHTTP protocol splits request identity from content across two parties. This CLI approach offers stateless privacy for agents, shifting key management burden to.
- Models & Research
Kimi K3 on M1 Max: Bandwidth, Not Capacity, Limits Local MoE Inference
Running Kimi K3 on an M1 Max proves local MoE feasibility via expert offloading, but unified memory bandwidth caps throughput. Treat this as a prototyping tool, not a serving.
- Ethics, Policy & Safety
Why Written AI Policies Fail to Control Agent Behavior
HANDBOOK.md benchmark shows frontier agents follow 124-page policies on only 36.2% of trials. Governance requires runtime enforcement, not documentation.
- Agents & Frameworks
Why Multi-Agent LLM Delegation Concentrates Risk
Splitting tasks across planner and worker agents concentrates risk because principals cannot observe hidden actions. Base models defect. Teams must add verifiable execution.
- Models & Research
Kimi Linear Cuts KV Cache 75% but Recall Remains the Binding Constraint
Kimi Linear cuts KV cache 75% and boosts decode 6x for 1M-token contexts, but recall stays the binding constraint. Route memory-bound jobs only after benchmarking in-context.
- Ethics, Policy & Safety
Why Vendor Model Cards Fail Clinical Ethics Procurement
PrinciplismQA exposes a gap generic leaderboards miss. Hospital compliance teams must build domain-specific ethics evals to test autonomy versus beneficence conflicts before.
- Agents & Frameworks
Harness vs Scaffold: Why Claude Code and LangGraph Are Not Interchangeable
A 2026 study shows harness choice swings coding-agent token costs by 40x while pass rates barely move. Misplacing tool gating in the scaffold layer creates integration debt.
- Developer Tools
Fine-Tuning vs RAG for Internal APIs: StarCoder2 Constraints
StarCoder2's 4,096-token sliding attention limits RAG for internal APIs. Fine-tuning encodes proprietary signatures in weights but requires a permanent dataset curation loop.
- Agents & Frameworks
MCP Tool Discovery Moves From Hardcoded Config to Runtime Agent Search
Runtime tool discovery moves the MCP trust boundary from manifest review to per-call authorization, turning a lockfile choke point into unbounded binding decisions.
- Infrastructure & Runtime
Calibrated LLM Monitoring: Conformal Prediction with Drift Detection
A cloud-native architecture packages conformal prediction as a sub-2ms microservice to replace raw logprob confidence with finite-sample coverage guarantees, gated by.
- Developer Tools
Mellum2 Unverified: Why MoE Active Parameters Matter More Than Total Size
Mellum2 specs lack sources. DeepSeek-Coder-V2 proves MoE economics. Local teams must demand active parameter counts before deploying coding models.
- Agents & Frameworks
Code-as-Action Agents Beat GAIA But Require Runtime Sandboxing
Hugging Face's code agent leads GAIA by treating code as the action space. This generalizes reasoning but shifts safety from prompt filtering to runtime containment, where.
- Ethics, Policy & Safety
TRIDENT Benchmark: LLM Safety Gaps in Finance, Medicine, and Law
TRIDENT benchmarks LLM safety in finance, medicine, and law using professional ethics codes. Results show specialized models often fail subtle ethical tests that generalists.
- Models & Research
Why Chat Leaderboards Do Not Predict Image Quality
A July 22 visual test comparing GPT-5.6, Claude, Gemini, and Grok highlights a structural gap: chat models lack dedicated image backends. Teams must route by subtask using.
- Developer Tools
PyPI Wheel Reproducibility: 15% Byte-Identical, 79% Source-Equivalent
Only 15.4% of PyPI wheels rebuild byte-identically from source. A new preprint measures 12,180 releases to expose the gap between pip-audit trust and actual source.
- Models & Research
Kimi K3 Procurement: Governance Review Over Phantom Government Assessments
Moonshot AI's Kimi K3 release triggers governance review for regulated teams. Beijing jurisdiction, Anthropic accusations, and missing government assessment require internal.
- Ethics, Policy & Safety
EU AI Act Traceability: Why ML Pipelines Fail Conformity Assessment
ML teams shipping high-risk systems to the EU face rejection if they cannot produce lifecycle-wide traceability artifacts. A May 2026 survey shows these mechanisms are often.
- Infrastructure & Runtime
Cloudflare AI Crawler Controls: Block, Charge, or Allow Bots Per Route
Cloudflare ships per-route AI crawler controls for Search, Agent, and Training bots. Site owners must model pay-per-crawl yield against ad inventory and measure Agent.
- Agents & Frameworks
Why Agent Security Tests Must Audit Full Trajectories, Not Single Turns
ASEval shows agent risk triggers jump from 22.9% to 47.4% when testing full multi-step trajectories instead of single turns. Vendors certifying agents on single-turn.
- Agents & Frameworks
x402 Per-Call Payments: Agent Wallet Custody and Replay Risks
x402 revives HTTP 402 for per-call stablecoin payments. The standard shifts wallet custody and replay protection burdens to agent runtimes, fitting brokered workflows better.
- Models & Research
DeepSeek Compute Leak: Why Open-Weight Routing Needs a Swap Path
A leaked transcript claims DeepSeek faces a compute gap, but V4 shipped in April. Teams should treat this as a resilience prompt to pin Qwen as a swap-in spine, not panic.
- Models & Research
Context Ordering Beats Window Size for Long-Context Agents
ARBIGRAPH shows tool agents lose 33.3% accuracy on dependent chains despite fitting context. Context ordering, truncation, and caching now beat raw window size for agent.
