https://network-ai.org/journal/ Release notes and shipped version updates for Network-AI. en-us Tue, 29 Sep 2026 16:22:00 GMT https://network-ai.org/blog/network-ai-v5-15-4/ https://network-ai.org/blog/network-ai-v5-15-4/ Tue, 29 Sep 2026 16:22:00 GMT ClawHub skill: project-context memory poisoning fixed (scripts/contextmanager.py, A.I.G T02). init/update now validate and reject unsafe values before writing, with per-section type checks. Every field and key is scanned https://network-ai.org/blog/network-ai-v5-15-3/ https://network-ai.org/blog/network-ai-v5-15-3/ Sun, 27 Sep 2026 21:56:01 GMT Public hardcoded orchestrator token replaced with per-instance secrets. The orchestrator's full-access blackboard identity used the constant 'system-orchestrator-token', also hardcoded in bin/mcp-server.ts, TaskDecompose https://network-ai.org/blog/network-ai-v5-15-2/ https://network-ai.org/blog/network-ai-v5-15-2/ Sun, 27 Sep 2026 20:39:30 GMT GHSA-9p2w-prp8-5722 (High) - ClaudeHookBridge deny patterns now inspect every string field of toolinput (e.g. Write.content, Edit.newstring, MCP tool arguments), not just the first candidate field. Over-nested or oversiz https://network-ai.org/blog/network-ai-v5-15-1/ https://network-ai.org/blog/network-ai-v5-15-1/ Tue, 28 Jul 2026 19:03:24 GMT Fixes two high-severity vulnerabilities reported against the v5.14.0/v5.15.0 ecosystem-expansion and context-composer features shipped in the last two releases. Both are validator/executor representation mismatches (CWE- https://network-ai.org/blog/network-ai-v5-15-0/ https://network-ai.org/blog/network-ai-v5-15-0/ Mon, 06 Jul 2026 21:15:25 GMT Agents have large context windows, but their effective reasoning window is smaller: irrelevant, stale, or noisy context degrades output quality long before the hard token limit ("context rot"). This release makes curated https://network-ai.org/blog/network-ai-v5-14-0/ https://network-ai.org/blog/network-ai-v5-14-0/ Sun, 05 Jul 2026 20:21:28 GMT Making Network-AI the neutral coordination layer for every major agent ecosystem: Claude, OpenAI, Gemini, and OpenClaw. https://network-ai.org/blog/network-ai-v5-13-4/ https://network-ai.org/blog/network-ai-v5-13-4/ Sun, 05 Jul 2026 19:03:24 GMT The v5.12.2 secret / checkAuth() gate covered only the two mutating routes (POST /approve, POST /deny), leaving GET /, GET /stats, GET /sse, and GET /:id unauthenticated even when a secret was configured, disclosing queu https://network-ai.org/blog/network-ai-v5-13-3/ https://network-ai.org/blog/network-ai-v5-13-3/ Fri, 26 Jun 2026 20:02:30 GMT SkillSpector YARA agentskillmcptoolpoisoningmetadata — Reworded the audit-log privacy note in SKILL.md that was triggering the exfiltration sub-rule. No functional change; VirusTotal 64/64 clean.