Habr•October 11, 2026•🇷🇺Translated from Russian

Developer Spends $9,000 on AI Agents to Build Crossweft Tool for Enforcing Multi-Language Component Agreements

A developer building a retouching application and Photoshop plugin with a local neural network spent more than $9,000 on AI agent subscriptions while constructing ten security layers across multiple processes and languages. The layers included a protected channel between plugin and system service, device registration, timestamped artifact signing, build attestation, update channel, model encryption, and watermarking. These components, written in C++ and Go, required 54 inter-component seams that compilers could not check because the sides used different languages and lived in separate processes.

The developer identified these seams as points where two or more sides must exactly match: service names, protocol versions, DTO fields, environment variables, and duplicated business logic such as price rounding. After three concrete failures—an identifier defined in three nearly identical functions, a dependency resolved from two different directories, and a security validator that scanned zero files for weeks—the need for automated enforcement became clear. AI agents proved especially prone to leaving one side of a seam unchanged because they performed only local edits and defined “done” as passing the checks visible in their current context.

Crossweft solves the problem by requiring every seam to declare its enforcement method in a JSON map placed next to the code. Each connection is anchored to literal strings so that code movement immediately invalidates stale descriptions. Enforcement types include shared-code, generated, schema-tests, duplicated, and convention. Duplicated and convention seams receive dedicated guards that read both sides of the seam; a guard that reads only one side is rejected.

Three simple guard types cover most cases: join verifies identical values, set verifies identical sets of fields or constants, and pair verifies duplicated logic blocks marked with crossweft:begin and crossweft:end tags. Larger architectural rules check that every server route appears in the map, that every client call is explained by a documented link, and that data-flow invariants hold. Thirteen custom plugins extend the system, including one that reconciles 916 Go SQL calls against 161 tables and 276 triggers.

Agent integration uses three hooks: a session-start summary of the seam map, an after-edit reminder naming the opposite file and required guard, and a pre-finish crossweft check that blocks completion until all seams agree. When an agent changes only one side, the attest command refuses to accept the change unless the developer explicitly confirms the other side remains unchanged. Findings that cannot be fixed immediately are recorded with owners and next steps; once resolved they automatically invalidate the record.

After the seam map and guards were added in late September, the installer that had previously failed on clean virtual machines succeeded for the first time. The number of active validators rose from four to fifty, each required to output SCANNED counts and pass self-tests. The open-source release includes the crossweft command, plugins for Claude Code and Codex, and an MCP server so any compatible agent can treat seam verification as a read-only tool.

Related articles

Habr•AI Security

Docker Rebrands Around AI Agents with 'Run on Trust' and New Sandbox Tools

Docker has updated its branding to emphasize security for AI agents, introducing the slogan 'Trust Docker for the agents you don’t' and the tagline 'build, ship, and run on trust'. The company highlighted products such as Sandboxes, AI Governance, MCP Enterprise Gateway, and Hardened Images, although most of these features were released or discussed earlier in 2025. Sandboxes run agents inside microVMs with separate kernels using KVM or Windows Hypervisor Platform, allowing isolated execution of tools like Claude Code. AI Governance provides centralized policy management and audit logging for enterprise teams, while MCP Enterprise Gateway handles SSO and credential injection for MCP servers. Hardened Images remain available under Apache 2.0 with optional paid FIPS and support tiers. The shift follows the appointment of Don Johnson, formerly of Oracle, as CEO and reflects Docker’s attempt to address new security concerns around autonomous AI agents rather than traditional container workloads.

Habr•AI Security

AI Agents Exploit PaperCut Flaw, Compromising 395 Organizations Across 48 Countries

PaperCut released a security bulletin on 27 August 2026. GreyNoise observed the first real-world code execution just four days later, with the operator moving from initial access to execution in under four hours. After the mass campaign launched, 11 organizations were compromised in only 26 seconds. In total, at least 440 PaperCut NG/MF instances belonging to 395 organizations were breached. The incident is detailed in the November report by Google Threat Intelligence Group and Mandiant, which examines the role of AI agents in the attacks. GTIG also discovered an automated exploit_agent system capable of handling exploitation without constant human input. Additional cases involved LLMjacking and unauthorized configuration changes in Claude Code repositories at cryptocurrency firms.

安全客•AI Security

Google Threat Intelligence Group Warns of Fully Autonomous AI Cyberattacks Within Six Months

Google's Threat Intelligence Group (GTIG) has observed the first real-world instances of attackers using multi-agent AI frameworks to autonomously discover vulnerabilities, generate exploits, and compromise targets without human intervention. The discovery came from an exposed server running a framework called exploit_agent that analyzed software update logs from popular web applications to identify and exploit weaknesses. Four previously defensive tools—ARTEX, CAIRN, CyberStrikeAI, and HexStrike AI—have been repurposed for offensive operations, with documented cases targeting French media, U.S. aerospace, and cryptocurrency firms. GTIG projects that low-complexity automated attacks will become common within three months, while sophisticated fully autonomous operations are expected within six months. The report highlights how these systems compress patch windows, overwhelm security teams with AI-generated proof-of-concepts, and turn detailed vendor advisories into attack playbooks. Defenders are urged to minimize public exposure, accelerate patching cycles to weekly, and tightly control API keys to prevent LLMjacking.

Habr•AI Security

OpenAI Rolls Out KYC, Hardware Keys and Tiered Access for Cybersecurity Tasks in GPT-6.2

OpenAI has introduced stricter access controls for GPT-6.2 Astra and Sol models when users attempt cybersecurity-related tasks. Certain legitimate security research and testing activities now trigger automatic blocks unless the user obtains Daybreak trusted access. The program requires paid subscriptions, FIDO2 hardware keys, Advanced Account Security, and identity verification through Persona. Developers report that even after meeting preliminary requirements, access to reduced refusal rates is not guaranteed and may require separate organizational approval. The changes stem from the models reaching Critical capability levels under OpenAI’s Preparedness Framework for dual-use cybersecurity skills. Similar verification processes previously led to account blocks for users of competing services such as Anthropic’s Claude.