Autonomous AI Just Filed Its First Incident Reports
In ten days, OpenAI and Anthropic both disclosed that their own models breached real companies during safety testing. Nobody was attacked. These were evals.
POSTS
All posts on software engineering, AI agents, Claude Code, test automation, and Python by Mor Dabastany.
26
posts
(3 pages)
HTTP Finally Got a QUERY Method, and It Fixes a 20-Year-Old Hack
RFC 10008 gives HTTP a safe, idempotent, cacheable method that carries a body. Here's why the QUERY method matters and what it kills.
The Token Burning Bubble: Inside the Collapse of Silicon Valley's Most Expensive Metric
For a year, burning AI tokens became a proxy for ambition. Then the bills came due. A look at tokenmaxxing, Goodhart's Law, and the shift to token efficiency.
Why AI Agents Fail at Cross-Repo Code Design - And How to Fix It
AI agents write code fast but design badly across many repos. The fix isn't a smarter model, it's letting the agent query the codebase instead of walking it.
Claude Code /goal: Set the Finish Line, Walk Away
The /goal command sets a completion condition and Claude works until a separate evaluator says it's done. How to write goals that actually finish.
Launchpad - Feature Delivery as a Four-Step Flow Inside Claude Code
Launchpad is a Claude Code plugin that turns feature work into four guided slash commands - brain, architect, implement, test - with state that survives /clear.
Poor Man's Agent - Claude Code Headless, No API Key Required
Stop paying API rates to call your own assistant. claude -p turns a Claude Code subscription into a scriptable agent that already knows your skills.
Claude Code on a Token Diet - Trim the Bill, Not the Brain
A short, honest playbook for cutting Claude Code costs without switching to a worse tool.
BMad Builder - Build Agents You Can Trust
BMad Builder generates the agents, skills, and workflows your team would otherwise hand-write. Here is the short version.
Nobody Tests Their AI Agent Skills. Here's Why That's a Problem (and How to Fix It)
47,000+ agent skills across 6,300+ repos. Almost none tested beyond a vibe check. Here's a 4-layer testing architecture that fixes that.