OSSbenchflow-ai/skillsbench Agent Skills BenchmarkGitHub·high signalXBlueskyLinkedInCopy linkFirst standardized benchmark for agent skills. Key finding: curated skills boost 16.2% but self-generated skills provide zero benefit.SourceSource pageGitHub↳ Follow the threadStack layer / Threat patternRAGFlow 0.27.2 rewrites its Agentic RAG retrieval framework and patches a starlette CVEGitHubStack layer / ContrastPydantic AI 2.42.0 adds a first-class provider for GitHub Copilot's OpenAI-compatible APIGitHubStack layer / Update threadCline Desktop 0.0.25 lets Claude Code and Codex CLI providers start sessions with no API key, and caps Codex models at real backend budgetsGitHub ReleasesStack layer / Threat pattern16% of 3,171 public agent-harness setups carry a confirmed security defect, and 3.8% ship a skill that pre-approves your shellarXivStack layer / Follow-up threadColibrì runs 744B to 2.8T MoE models on consumer hardware in pure C by streaming experts off diskGitHub TrendingThreat patternAgentTrust Grades 50 MCP Servers A Through F and Ships the Scoring as a CI Gate With SARIF Outputeulogik/AgentTrust on GitHub (metadata verified via the GitHub API; single source, 2 stars)Stack layergentle-ai carries 919 open items, the largest backlog on today's boards, at 6,583 starsGitHub TrendingStack layerThree separate Claude Code bugs were silently destroying prompt-cache reuse for subagentsGitHub