Prompt injection is a data-plane problem: move the boundary from the model to the tool call.

The author argues that prompt injection shouldn't be solved by making models smarter; instead, just as SQL injection is handled with parameterized queries, the boundary should be drawn where the agent executes actions.

Skills Are Not Tools: Why I Gave My Coding Agent 11 MCP Servers and It Fell Apart

In February I hooked up 11 MCP Servers to my coding agent. Just the tool list in an empty session ate up 34,000 tokens, with Datadog alone contributing over two hundred tools. The agent got slower and kept picking the wrong tools.

Coding agents are like steroids: what bodybuilding can teach us about the risks of AI coding tools

Some developers have drawn the analogy that coding agents are like steroids in bodybuilding: they let beginners achieve with ease what used to take enormous effort, and let experienced professionals produce workloads that were previously impossible. But these code-generation agents carry medium- and long-term risks, and amateurs and beginners should wait until they hit their own "natural limit" before considering using them.

Matt Pocock@mattpocockukAI score4040

After talking with poteto, Matt Pocock argued that in the AI era we should use abstractions more, backed by strict lint rules that narrow the agent's design space and keep it making good decisions. High-leverage abstractions let you do more with less code, which saves tokens; and with agents around, cleaning up the damage from a bad abstraction is far cheaper. He says this runs against the popular belief that "agents just want to read raw code," and calls for being bolder about designing abstractions.

Addy Osmani on engineering methods for bringing AI agents into legacy codebases

Addy Osmani suggests that when introducing AI agents into brownfield codebases, you should first make hidden constraints visible and make cheap changes trustworthy. He recommends dividing code into green, yellow, and red zones: green zones with solid tests let agents move in small, fast steps; yellow zones require writing characterization tests first; and red zones involving sensitive logic like authentication, billing, and permissions must have humans involved step by step. The zones are drawn by hand, and a yellow zone can only be upgraded to green once characterization tests exist and the module owner has reviewed the first batch of changes.

An agent platform built for inventing agents: decoupling capability interfaces from their implementations

Author Ryan Lopopolo argues that an agent is a parameterized program built on top of a set of capabilities: models and configurations, reasoning and tool-call loops, computers, disks, context, Skills, tools, connectors, runtimes, network policies, identity, IAM, guardrails, I/O channels, and system prompts.

Addy Osmani on software factories: the visible factory and the hidden factory, where validation is the bottleneck

Addy Osmani proposes that a software factory has three layers—loop, harness, and factory. The factory isn’t a smarter agent; it’s multiple loops with harnesses feeding into a single review gate, with humans controlling the outer loop.

Code Red needs a maintenance loop: use Codex /goal to turn emergency fixes into ongoing operations

Drawing on his experience during Stripe’s first code yellow, the author points out that after most code reds, all that’s left is a post-mortem and exhausted engineers, while the metrics go back to being unowned. He argues that a code red should leave behind a maintenance loop, and that OpenAI Codex’s /goal command can turn a one-off coding request into an ongoing objective with clear completion criteria, letting a persistent cluster of agents continuously watch metrics, generate interventions, and request human review.