A developer discovered that their CLI tool llm-council, which chains multiple language models, was vulnerable to prompt injection because the fence markers used to delimit untrusted content were fixed…
A developer's self-hosted observability stack on Railway silently failed for six days because Prometheus stopped writing to disk due to a full volume, yet queries continued from memory, masking the is…
A developer building CodeBots Arena, an AI vs AI code battle platform, fixed a CI bug where tests passed locally but failed in GitHub Actions due to Python's module search path. The developer updated …
Addy Osmani, a 14-year Google veteran who led Chrome DevTools and Core Web Vitals, discusses his transition to AI engineering in the latest Pragmatic Engineer podcast, revealing that he built a web br…
A developer's guide to extending Claude Code with MCP servers, subagents, and hooks details how to connect external services, manage scopes, and create custom subagents. The guide covers adding MCP se…
A bug in Sentry's JavaScript SDK double-counts tool calls in streaming mode for Google's Gemini integration, recording each invocation twice with inconsistent parameter keys (`args` vs. `arguments`). …
A developer discovered that a security hook guarding an AI coding agent silently stopped protecting files with non-ASCII paths due to a one-line encoding bug. The hook, which reads JSON from stdin, us…
Developer Abbas Mirza built Aether Canvas during OpenAI Build Week and discovered a silent data-loss bug in its workspace index persistence. The bug allowed two concurrent operations to overwrite each…
At DEF CON 34, Tenet Security demonstrated Ghostjacking, an indirect prompt injection attack that succeeds 90% of the time against Claude Code under Cloudflare's default recommended configuration, all…
A solo developer found that a duplicate function definition in a 16,000-line Flask app silently broke GIF trimming, causing the tool to convert entire videos instead of selected clips. The bug, caused…
A developer building an autonomous AI agent named Hermes Commander encountered cascading failures when the agent attempted to invoke a web search tool that was not registered, leading to a blind fallb…
A developer wired Sentry error monitoring into the LLM pipeline of TextStack, an open-source .NET reader, after discovering that a user's PDF fell onto a CPU-only Ollama container, causing 390% CPU us…
A developer building adapters for a protocol indexer discovered that an end-to-end test with a tolerance for chain reverts was masking a parser bug, causing the suite to pass despite the parser failin…
Tenet Security unveiled the Ghostjacking attack at DEFCON 34 on August 9, demonstrating a 90% success rate against Claude Code under Cloudflare's default settings by poisoning logs from platforms like…
Tenet Security Threat Labs disclosed a new attack technique called Ghostjacking that turns logs and alerts into commands to take over AI agents. The attack embeds natural language commands into WAF lo…
A developer named Sarvar discovered a hidden retry in a five-agent pipeline using Sentry's span hierarchy, where one agent took 22.6 seconds due to retry logic while others took 5 seconds. The case hi…
Charlie, an AI engineering agent, now provides legible work records that answer five questions—lifecycle, task tree, evidence trail, and current artifact—without exposing private chain-of-thought, acc…
A developer fixed a Godot 4 bug in their game Nocturne Vania where a stale global class cache caused parse errors on local startup but passed CI. The issue occurred because CI imported the project bef…
Sentry traces revealed a page load time of 4 seconds, with TTFB alone at 2,068ms, prompting a developer to optimize their Nginx cache and move API calls to the backend. Using Sentry's AI assistant See…
Beacon, a self-hosted error tracking and LLM observability tool from developer Tboworst, ingests errors from any service, groups them by root cause, and displays them in a live terminal dashboard or b…