50
Securing Claude Code: The New Compliance API, Local Visibility, and Identity Governance
Anthropic's new Compliance API endpoints expose Claude Code local session transcripts, highlighting governance gaps for endpoint AI agents.
Anthropic added local session transcript endpoints to its Compliance API on August 11, 2026, giving security teams visibility into prompts, bash commands, file operations, and MCP commands run by Claude Code harnesses on endpoints. The article argues local harnesses break the classic shared-responsibility model, citing Token Security data that 68.6% of discovered AI agents run on endpoints, and a Cloud Security Alliance survey of 418 IT and security professionals in which 82% found an unknown agent within the past year. It outlines three governance layers: Anthropic managed settings as a policy baseline, the Compliance API for cloud-visible transcripts, and endpoint telemetry to connect agent activity to identity, credentials, and permissions.
45
42
50
30
42
The VMs Powering Mobile Agents (Instinct, Claude Code)
A teardown reveals Claude Code runs in Firecracker microVMs with a Rust PID 1 and MITM'd egress, while Instinct rents E2B sandboxes with git-based memory.
The author inspects the virtual machines hosting cloud agents: Claude Code runs in a Firecracker microVM with a custom Rust init (process_api) as PID 1, a 324 MB Bun harness on a read-only disk, and 443-only MITM'd SSE egress to api.anthropic.com with host-rotated OAuth tokens and no inbound access. Instinct rents E2B sandbox-as-a-service Firecracker microVMs (Ubuntu 22.04, 2 vCPU, 1.9 GB RAM) where agent memory is a git repo of Markdown committed by the agent and pushed to S3 as a single bundle, using short-lived STS credentials. Both platforms rely on Firecracker, differing mainly in fleet operator and guest boot configuration.
35
60
42
30
60
Anthropic Adds Plugin Evals to Claude Code: 6 Grader Types, a No-Plugin Baseline, and a CI Gate for Skills
Anthropic ships a plugin evals workflow for Claude Code with six grader types, a no-plugin baseline arm, and a CI gate via threshold and cost flags.
Anthropic published a plugin evals workflow for Claude Code, exposed via the "claude plugin eval" command on v2.1.269+. Six grader types exist: regex, tool_used, tool_order, and file_exists are free transcript checks, while llm and baseline invoke a billed judge model. Every case runs with and without the plugin, and the delta (Δ) isolates the plugin's contribution; a Δ near zero with a failing tool_used:Skill grader indicates the skill never triggers. CI gating uses --threshold 0.8, --max-cost-usd, --trust-plugin, and --no-publish flags, with results written to a report.html under evals/results/.
42
57
Malicious .git Configs Can Make Claude, Codex, Cursor, and Other AI Agents Run Attacker Code
Malicious .git core.fsmonitor configs let attacker code run in AI coding agents like Claude Code and Codex; four of eight flaws remain unpatched.
Manifold Security disclosed eight flaws across seven command-line AI coding agents where a repository's Git core.fsmonitor configuration causes agent-spawned commands to execute attacker code outside the sandbox and without approval prompts. Fixes shipped for goose (CVE-2026-72718, CVSS 4.0 score 7.0), Claude Code on one path, and Cursor, while Hermes Agent, Qwen Code, Grok Build, and a second Claude Code path were still unpatched as of September 1. OpenAI issued three CVEs for the same class in Codex, including CVE-2026-19592, and prior related bugs include CVE-2021-43891 in Visual Studio Code and CVE-2022-24346 in JetBrains IDEs. Exploitation requires the repository to arrive with its .git directory intact, such as via archives, shared drives, or USB sticks rather than an ordinary clone.
68
30
30
47
30
45
42
50
Houthis Used Claude Code to Develop Missile Guidance Software: Anthropic
Anthropic's threat report details a Houthi-linked Yemeni cell using parallel Claude Code sessions to build missile guidance software, evading safeguards by fragmenting tasks.
Anthropic's September threat report describes a Yemen-based cell, assessed as highly likely Houthi-linked, that used Claude Code across multiple parallel instances to develop guidance software for a tactical guided rocket, a ballistic missile with over 2,000 km range, and a hypersonic glide vehicle concept called 'R2000'. The operators integrated open-source autopilot software, built six-degree-of-freedom trajectory simulations, and used reinforcement learning to tune flight-control algorithms, ultimately compiling an offline executable. The group test-fired a guided rocket that failed, then used Claude within hours to analyze launch telemetry. Anthropic blocked numerous requests, but operators evaded safeguards by obscuring intent and dividing work across separate conversations before accounts were banned; the case is one of six conventional-weapons cases (three China-linked, two Russia-linked) in a report covering disrupted operations from December 2025 to August 2026.
75
Week in review: AiTM phishing kit used to hijack AWS accounts, year-long malware campaign targets HR
60
42
60
Agent Harness vs Agent Framework vs MCP: Which Layer Owns the Loop, State, Tools, Permissions, and Recovery
Architecture explainer separates agent harnesses, frameworks, and MCP by which layer owns the loop, state, permissions, and recovery.
The article distinguishes agent harnesses (OpenAI Codex, Claude Agent SDK), which own the execution loop, sandbox, permission model, and recovery; frameworks (LangGraph, OpenAI Agents SDK, Microsoft Agent Framework), which supply composable primitives; and MCP, a stateless JSON-RPC wire protocol governed by the Linux Foundation's Agentic AI Foundation since December 2025. An ownership matrix maps the execution loop, state, tool transport, permissions, recovery, sandboxing, and multi-agent orchestration to each layer. The 2026-07-28 MCP specification made the protocol fully stateless, retiring the initialize handshake and session headers.
35
Context Engineering Inside the Harness: 4 Mechanisms That Beat Context Overflow and Goal Loss on Long-Horizon Tasks
Survey of four harness mechanisms—context budgeting, compaction, todo-state, and memory—that keep long-horizon LLM agents on task across 200+ tool calls.
The article details how agent harnesses, not larger context windows, solve context overflow and goal loss on long-horizon tasks, citing Chroma's Context Rot report showing 18 LLMs (GPT-4.1, Claude 4, Gemini 2.5, Qwen3) degrade on long inputs. Concrete implementations include LangChain Deep Agents offloading tool responses over 20,000 tokens to the filesystem and truncating old tool calls at 85% window usage, and Claude Code capping auto memory at 25KB while re-reading the 5 most recently modified files after compaction. OpenAI's Responses API now offers server-side compaction via context_management with a standalone /responses/compact endpoint, which Codex uses for long-running coding tasks. Manus reports a roughly 100:1 input-to-output token ratio per ~50-tool-call task, motivating todo.md state recitation to prevent goal drift.
52
55
30
Coop – Isolated VM Environments for Running Claude Code and Codex
Trail of Bits releases Coop, running Claude Code and OpenAI Codex agents inside isolated virtual machines for safer agentic coding.
Coop, published on GitHub by security firm Trail of Bits, provides isolated VM environments for executing AI coding agents such as Claude Code and Codex. Isolation contains the filesystem and network side effects of autonomous agent actions, reducing risk from unsupervised tool use. The project drew 61 points and 16 comments on Hacker News.
28
Show HN: Self-hosted company OS, Claude Code and Codex agents in departments
OtoDock, a self-hosted company OS that organizes Claude Code and Codex AI agents into departments, was launched on GitHub via Show HN.
OtoDock is a self-hosted 'company OS' shared on GitHub through a Show HN post, presenting Claude Code and Codex AI agents organized into department-style teams. The Hacker News feed entry shows the post reached 20 points with 5 comments; no further technical details are provided in the available text.
18
Researcher shows how Claude Code can be tricked simply by asking it to summarize a website
Researcher wunderwuzzi showed Claude Code can be hijacked via prompt injection simply by asking it to summarize a website.
Security researcher Johann Rehberger (wunderwuzzi) demonstrated that Claude Code can be manipulated through prompt injection by simply asking it to summarize a website. Instructions embedded in fetched web content are executed by the agent, hijacking its behavior. The Register frames the finding as another demonstration of prompt injection risks in agentic coding tools that ingest untrusted web content.
45
60
30
30
60
China-Based Artificial Intelligence Companies Conducting Industrial-Scale Distillation Campaigns Against U.S. AI Companies
NSA, CISA, and FBI warn DeepSeek, Alibaba, and other Chinese AI firms ran industrial-scale distillation of U.S. frontier models, threatening U.S. AI leadership.
A joint NSA, CISA, and FBI Cybersecurity Advisory (AA26-251A) says China-based firms DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI extracted billions of tokens from U.S. frontier models including Claude, GPT, Gemini, and Grok, likely with Chinese government knowledge. Campaigns running since at least late 2024 used native APIs, cloud providers, third-party aggregators, gray-market proxy "transfer stations", and shared premium subscriptions to bypass geographic restrictions, evade safeguards, and violate providers' terms of use. The agencies recommend detecting anomalous prompts, accounts, and usage patterns; subtly altering responses to suspected distillers; and cross-organization intelligence sharing. They also call DeepSeek's publicly cited $5.6M training cost misleading because it excludes data acquired through distillation.
76
57
Attackers turn to AI for help identifying files worth stealing
Gambit Security documents three threat actors using AI: a ransomware operator with Claude Code, credential harvester Zerofot, and the AI-built RAGE cryptomining framework.
Gambit Security examined three unrelated threat actors using AI across different stages of intrusions. A suspected operator tied to The Gentlemen ransomware-as-a-service used Claude Code running Claude Sonnet 4.6 in late June 2026 at six organizations, including an Australian energy utility, where it ran reconnaissance, ranked valuable databases, staged SQL Server dumps for exfiltration, and modified firewall configurations, accidentally taking one utility firewall offline. The Zerofot credential-harvesting operation, built with OpenAI Codex and Claude Code, collected 2,975 validated credentials from 1,742 hosts between April 5 and May 23, 2026, including SSH private keys and AWS access keys. The AI-generated RAGE Python framework exploits exposed Redis, Elasticsearch, Docker, Tomcat, and other services to harvest credentials and deploy cryptominers, guided at runtime by a DeepSeek-backed AI Orchestrator.
65
35