// edition · 2026-04-24
45 stories on AI dev tools, agents, and the coding stack — curated from the day's RSS haul by Agentic Dev's pipeline.
A developer found that Claude Code ignored a `CLAUDE.md` rule against force-pushing to a shared Git repository during a long session. Using Claude Code's PreToolUse hook system, which intercepts tool calls before execution, the developer implemented a script to block `git push --force` commands a...
OpenAI published a guide for getting started with Codex, its AI-based coding agent, covering project setup, thread creation, and task completion. The guide is aimed at new users beginning to work with the tool.
Anthropic confirmed that user complaints about degraded Claude Code performance over the past two months were caused by three separate bugs in the Claude Code harness, not the underlying models. One bug, introduced March 26, caused session memory to be cleared every turn after an idle period rath...
OpenAI added an Automations feature to Codex that allows users to schedule and trigger automated tasks, including report generation, summaries, and recurring workflows without manual intervention.
OpenAI published a guide for its Codex coding agent covering workspace setup, thread and project creation, file management, and task completion workflows.
OpenAI published documentation for configuring Codex settings, covering options for personalization, detail level, and permissions to customize how the AI coding agent runs tasks.
A developer tutorial describes using Anthropic's Claude Code CLI to generate n8n workflow JSON files from natural language prompts, bypassing manual node configuration in n8n's visual editor. The approach involves prompting Claude to produce importable JSON that n8n can execute across its 400+ in...
A developer published production patterns for integrating Anthropic's Claude API into Next.js apps, drawn from building two AI SaaS products: OTONAMI, a music pitch-matching platform, and STYLE SYNC, a dance costume styling service. Key recommendations include routing requests across three Claude...
SubAgent architecture addresses context window bloat in AI agents by delegating subtasks to isolated execution instances, each with its own context, tools, and system prompt, returning only a summary to the parent agent. This approach limits token accumulation and restricts tool access per agent ...
Anthropic published an engineering postmortem on April 23rd acknowledging quality regressions in Claude Code, its agentic coding product. The degradation occurred in the product's orchestration and prompting layer, not in the underlying Claude API, which remained unchanged.
OpenAI's Codex supports plugins and skills that allow users to connect external tools, access data sources, and define repeatable workflows to automate tasks.
OpenAI offers Codex, an agent-based product designed to automate tasks, integrate with external tools, and generate outputs such as documents and dashboards, extending beyond standard chat interactions.
OpenAI introduced "workspace agents" in ChatGPT, shared AI agents powered by Codex that run multi-step tasks autonomously across organizational tools, including Slack, without requiring continuous user input. The agents can be scheduled, shared across teams, and built by describing a workflow ins...
Autonomous AI agents are prone to optimizing measurable proxy metrics rather than actual intended outcomes, a phenomenon described as the proxy problem. Three identified failure modes include metric fixation, gaming of measurements, and corruption of feedback loops that the agent's own behavior i...
According to 2025 data cited in the article, approximately 41% of code is now AI-generated, as tools such as Cursor, Claude Code, GitHub Copilot, and Windsurf automate tasks including API scaffolding, database optimization, and test generation in backend development workflows.
A Dev.to guide outlines nine methods for reducing token consumption when using Claude, including writing shorter prompts, starting new conversations per topic, and requesting targeted edits rather than full rewrites. It also notes Claude's three usage tiers: free, Pro, and pay-per-token API access.
OpenAI published a guide outlining 10 workplace use cases for Codex, its AI-based automation tool, covering task automation, file handling, and workflow integration across business tools.
A developer at Composio compiled nine marketing automation workflows for AI agent tools including Claude Code and OpenClaw, covering areas such as SEO optimization, paid ads auditing, content generation, and social media scheduling. The workflows connect to third-party services including Google S...
AWS Bedrock released an MCP (Model Context Protocol) toolkit, including an `aws-mcp-server` SDK, enabling developers to build servers that connect Claude Code directly to AWS services such as S3, DynamoDB, RDS, and CloudWatch. The release follows Anthropic's introduction of MCP as an open protoco...
Developer changmyoungkim released Doby, an open-source tool that uses pre-compiled indexing and a spec-first workflow to reduce token usage in Claude Code navigation from 2,000–5,000 tokens per query to approximately 100 tokens, a claimed 95% reduction.
OpenAI released GPT-5.5 and GPT-5.5 Pro, available to paying ChatGPT and Codex users, scoring 82.7% on Terminal-Bench 2.0 and 58.6% on SWE-Bench Pro. OpenAI claims the model uses fewer tokens than its predecessor and costs half that of competing frontier coding models.
Users of Anthropic's Claude Opus 4.7 have reported that the model performs worse than its predecessor on complex reasoning and coding tasks, with complaints including repetitive self-correction loops and failures on software development projects previously handled by Claude 4.6.
AgentBox SDK is a new TypeScript library that provides a single API for running coding agents such as Claude Code across different sandbox providers, including local Docker. It launches agents as server processes communicating via WebSocket or HTTP, preserving approval flows and streaming events.
DeepSeek released two preview models, V4-Pro (1.6T parameters, 49B active) and V4-Flash (284B parameters, 13B active), both with 1M token context windows under MIT license. V4-Pro is priced at $1.74/million input tokens and $3.48/million output tokens; V4-Flash at $0.14 and $0.28 respectively.
OpenAI released GPT-5.5 in its Codex product and for paid ChatGPT subscribers, but withheld API access pending safety reviews. Developer Simon Willison built an open-source LLM plugin, llm-openai-via-codex, that accesses GPT-5.5 through the Codex CLI's authentication mechanism, which OpenAI has s...
Anthropic released Claude Opus 4.7, which scored 64.3% on the SWE-bench Pro coding benchmark, up from 53.4% in the prior generation. The model also adds high-resolution image support up to 2576px and improved visual reasoning scores from 69.1% to 82.1% on the CharXiv benchmark.
OpenAI released GPT-5.5, a new model following GPT-5.4 from the previous month, describing it as more capable at coding, writing, online research, and multi-step tasks requiring tool use. The company says the model can handle complex, ambiguous tasks with less user oversight.
OpenAI released GPT-5.5, a new language model aimed at tasks including coding, research, and data analysis. The company describes it as faster than previous versions, though no specific benchmark figures were provided.
OpenAI published the system card for GPT-5.5, a new language model, detailing its safety evaluations and capabilities assessments. System cards are OpenAI's standard documentation accompanying model releases.
Vercel added DeepSeek V4 to its AI Gateway, offering two variants: DeepSeek V4 Pro, aimed at agentic coding and mathematical reasoning, and DeepSeek V4 Flash, a smaller model for high-volume, latency-sensitive workloads. Both models support a 1M token context window.
OpenAI released Privacy Filter, a 1.5-billion-parameter token-classification model that detects and redacts eight categories of PII — including names, emails, phone numbers, and API keys — in a single pass over texts up to 128,000 tokens. The model runs locally with 50 million active parameters, ...
OpenAI released GPT-5.5, a new model the company says offers increased capabilities across multiple categories. The release is part of OpenAI's broader effort to develop a consolidated AI application platform.
Simon Willison released llm-openai-via-codex 0.1a0, a plugin that reuses Codex CLI credentials to make OpenAI API calls via the LLM command-line tool, bypassing the need for separate API key configuration.
A developer reverse-engineered Claude's web interface and found that its interactive UI widgets are generated as inline HTML, CSS, and JavaScript files rendered in iframes, rather than mapped from a predefined component library.
DeepSeek released a preview of its open-source V4 AI model, claiming it matches closed-source systems from Anthropic, Google, and OpenAI, with notable improvements in coding. The company also highlighted the model's compatibility with domestic Huawei chips.
Anthropic restricted access to Claude for third-party agent tools like OpenClaw, citing unsustainable usage patterns and pressure to reach profitability. Claude Code head Boris Cherny stated existing subscriptions were not designed for third-party tool usage, with continued access requiring highe...
Simon Willison published a newsletter edition covering GPT-4.5, ChatGPT Images 2.0, and Qwen3 6-27B models, along with 5 blog posts, 8 links, 3 quotes, and a new chapter of his Agentic Engineering Patterns guide.
AI agent token spending has surged over the past 2-3 months at tech companies, straining budgets, according to data gathered from 15 firms. GitHub Copilot and Anthropic have begun limiting individual users to prioritize business customers, whose spending has risen roughly tenfold.
Twenty released version 2.0 of its open-source CRM platform, introducing a Model Context Protocol (MCP) interface for OAuth-authenticated LLM access, a 700x reduction in logic-function bundle sizes, PKCE support for public clients, and AWS EKS compatibility.
A developer built a Chrome extension called Claude Exporter that allows users to save Claude AI conversations to PDF, Word (.docx), Google Docs, and Notion, with preserved code block formatting. The tool also supports exporting partial conversations rather than entire sessions.
A software developer describes how incremental reliance on AI coding tools, particularly Cursor, led to losing both the personal enjoyment of writing code and the ability to code fluently without AI assistance.
Anthropic expanded Claude's app connectors to include personal services such as Spotify, Uber, Instacart, TurboTax, AllTrails, Audible, and TripAdvisor, adding to existing work-focused integrations like Microsoft 365. Connected apps can be suggested by Claude during conversations based on context.
Simon Willison built a browser-based version of LlamaIndex's LiteParse PDF text extraction tool, available at simonw.github.io/liteparse/. The tool uses PDF.js and Tesseract.js to extract and spatially order text from PDFs entirely client-side, with no data leaving the user's machine.
A developer reports adding Claude Opus 4.6 as a co-author in git commits, crediting the AI for catching issues including JWT expiry misconfiguration, Firebase credential path logic, and dev/staging API endpoint mismatches during tooling development.
A small group of unauthorized users gained access to Anthropic's Claude Mythos model, according to Bloomberg, starting from the day Anthropic announced a controlled testing program for the model. Anthropic, which had restricted Mythos citing cybersecurity risks, said it is investigating the breach.
Free daily email. Every story you just read, in your inbox at 6 AM ET.