Coding Tools
GitHub Copilot Automations go team-shared
GitHub Changeloghigh confidence
Why it matters
- →Shifts AI workflow ownership from individuals to teams, preventing knowledge loss when staff leave.
- →Changes billing structure so organizations pay for shared agent usage rather than individual seat consumption.
- →Enables persistent, auditable automation for critical engineering tasks like bug reproduction and incident response.
GitHub has begun rolling out shared Copilot Automations, allowing teams to build and maintain AI-powered workflows collectively rather than individually. The update shifts billing for AI credits to the repository owner and attributes automation actions to the Copilot agent, ensuring workflows persist even if the creator leaves the organization. Available on Pro, Business, and Enterprise plans, this feature requires administrators to enable a specific policy for organization-owned repositories. This marks a significant step in treating AI agents as persistent team infrastructure rather than personal productivity tools.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1GitHub Copilot Integrates with Microsoft Teams
GitHub Changelog · August 21, 2026 · Same story
- 2GitHub Copilot Integrates with Slack in Public Preview
GitHub Changelog · August 21, 2026 · Same story
Follow this story
Open the full story →GitHub Copilot Automations go team-shared
3 developments
- Oct 28 · GitHub Changelog
GitHub Copilot Automations go team-shared (This article) - Oct 28 · GitHub Changelog
GitHub Copilot adds security extension for in-app code scanning↳ GitHub Copilot adds a security extension that pulls code scanning alerts directly into the app for in-editor explanation and fixes. - Oct 28 · GitHub Changelog
GitHub Enterprise bundles Copilot access
More in Coding Tools
Coding Toolscoding
Claude Code v2.1.316: Subagents and Gateway Budgets
Anthropic is tightening control over Claude Code's autonomous behavior with new environment variables that let you disable 'ultrathink' or cap subagent effort levels. For teams, the introduction of pooled group budgets on the gateway allows shared monthly spending limits across members, a practical fix for enterprise cost management. The release also patches critical bugs where shell commands would hang or crash on exit, stabilizing the developer experience significantly.
Claude Code Releases
Coding Toolscoding
llama.cpp b12022 adds ROCm 10 and Snapdragon support
This release quietly cements llama.cpp as the universal inference runtime by finally bringing ROCm 10.0 to Linux and Windows alongside CUDA 13.4, effectively closing the hardware gap for AMD users who were previously left behind. The inclusion of native Snapdragon support on Linux marks a significant step toward running local LLMs on ARM-based mobile chips without emulation overhead. While KleidiAI on Apple Silicon is disabled in this build, the broader expansion to diverse accelerators ensures that developers aren't locked into NVIDIA ecosystems anymore.
llama.cpp Releases
Coding Toolscoding
llama.cpp b12023 fixes shared prefix bug
This release addresses a specific inference artifact where decision prefixes were incorrectly bleeding into output embeddings. By capping shared prefixes at the first output position, llama.cpp ensures that option tokens remain distinct for parent and child tasks. This is a targeted fix for multi-step reasoning or voting workflows rather than a performance boost. It stabilizes complex generation patterns without changing the underlying model architecture.
llama.cpp Releases