Daily Agentic Field Watch - 2026-07-29 21:00 UTC
- GitHub Copilot Business and Enterprise now get a global default-enable policy for GA models. New models will be on by default after a 28-day config window, with a single org-level opt-out and explicit exclusions for open-weight / non-retained models. Sources: https://github.blog/changelog/2026-07-29-default-model-enablement-for-copilot-business-and-enterprise/ ; https://docs.github.com/en/enterprise-cloud@latest/copilot/how-tos/administer-copilot/manage-for-enterprise/manage-availability-of-default-models
- Microsoft Research pushed the agent-skill layer in two directions at once:
SkillOpt treats skill files as trainable parameters and reports best-or-tied-best results across 52 eval cells, while RESOURCE2SKILL distills tutorial videos, repos, articles, and reference artifacts into executable skills with provenance. Sources: https://www.microsoft.com/en-us/research/blog/skillopt-agent-skills-as-trainable-parameters/ ; https://www.microsoft.com/en-us/research/publication/resource2skill-distilling-executable-agent-skills-from-human-created-multimodal-resources/ ; https://www.linkedin.com/posts/pascalbiese_skillopt-exec-strategy-for-self-evolving-activity-7464627010712129536-fbpT ; https://www.linkedin.com/posts/eric-vyacheslav-156273169_you-can-now-turn-internet-tutorials-into-activity-7484502421139005440-dDB6
- A new Microsoft Foundry community write-up shows one agent ported across three runtimes: GitHub Copilot CLI + MCP, Microsoft Scout native tooling, and Foundry Local on-device NPU. The useful part is the architecture lesson: keep the behavioral core deterministic so the runtime can change without rewriting the agent; Scout wins on scheduling/memory and lower session cost, while Foundry Local trades reliability for offline/near-zero marginal cost. Source: https://techcommunity.microsoft.com/discussions/azure-ai-foundry-discussions/one-agent-three-runtimes-porting-a-csa-agent-to-microsoft-scout-and-foundry-loca/4541628
- Anthropic’s “Discovering cryptographic weaknesses with Claude” is getting real traction and real technical discussion: HN has a 200+ point thread, Reddit is discussing the HAWK and round-reduced AES findings, and the write-up says Claude Mythos found improved attacks rather than just implementation bugs. This is a good reminder that tool-using models are now meaningfully contributing to security research, for good and bad. Sources: https://www.anthropic.com/research/discovering-cryptographic-weaknesses ; https://news.ycombinator.com/item?id=49087091 ; https://www.reddit.com/r/crypto/comments/1v9dn75/discovering_cryptographic_weaknesses_with_claude/ ; https://www.reddit.com/r/cryptography/comments/1v99f4z/using_claude_mythos_preview_researchers_at/