Agentic autofix for code scanning is generally available
GitHub changelog says agentic autofix for code scanning alerts is generally available. It explores the codebase and proposes a fix for an alert.
Source: GitHub ChangelogFeed
What just shipped from tools in the directory. Every item has a date and a public source. Summaries are paraphrased. If a date is not public, it does not go in this feed.
GitHub changelog says agentic autofix for code scanning alerts is generally available. It explores the codebase and proposes a fix for an alert.
Source: GitHub ChangelogGitHub changelog lists code scanning AI Scan for pull requests as generally available, as a complement to CodeQL for changes in languages and frameworks CodeQL does not cover.
Source: GitHub ChangelogOpenAI release notes describe a GPT-Live-1.1 rollout in ChatGPT Voice, including a smaller Quip variant and additional voice options across several languages.
Source: OpenAI release notesGitHub changelog says Gemini 4 Argon is generally available in GitHub Copilot for complex coding, agent, and reasoning tasks.
Source: GitHub ChangelogOpenAI release notes describe a beta that suggests the next message in a Codex thread. Accepting a suggestion fills the box; it does not send the message on its own.
Source: OpenAI release notesOpenAI release notes say Codex in the ChatGPT desktop app can respond sooner when you send a follow-up to steer a running task.
Source: OpenAI release notesAnthropic platform release notes announce Claude Haiku 5.5 for high-volume, latency-sensitive work, with a 1M token context window. Code written for Haiku 4.5 can break; read the migration notes before you switch.
Source: Claude Platform release notesCursor changelog describes viewing and messaging local agents from the Cursor iOS app. The agent keeps running on your computer, which has to stay on and online.
Source: Cursor changelogAnthropic platform release notes announce Claude Sonnet 5.5 across the Claude API and several cloud platforms. The notes list request changes that can break code written for Claude Sonnet 5.
Source: Claude Platform release notesCursor changelog announces two bots for Teams and Enterprise: Rollouts reports deploy health per environment, and Security Review comments on exploitable bugs in a pull request. Bugbot remains the style and quality reviewer.
Source: Cursor changelogAnthropic platform release notes announce Claude Opus 5.5 for long-running agentic coding and knowledge work, with a 1M token context window by default. The notes say thinking cannot be turned off on this model.
Source: Claude Platform release notesCursor changelog describes Projects as a way to keep context on a larger body of work and delegate tasks to agents. The coordinator plans and reviews; it does not write the code itself. The beta started rolling out to all users.
Source: Cursor changelog