Search Results for "concrete"

...The goal of the project is to make it easy for AI assistants like Claude Code, OpenCode, Cursor, Codex, and others that support this open ecosystem to automatically apply best practices or perform concrete actions when a relevant user intent is detected. For example, some skills guide the agent in applying React and Next.js performance best practices, auditing UI and accessibility standards.

Inspect Petri is an open-source alignment auditing agent that lets researchers rapidly test concrete safety hypotheses against target models using realistic, multi-turn scenarios. Instead of building bespoke evals, Inspect Petri automatically generates audit environments from seed “special instructions,” orchestrates an auditor model to probe a target model, and simulates tool use and rollbacks to surface risky behaviors. Each interaction transcript is then scored by a judge model using a consistent rubric so results are comparable across runs and models. ...

...Multi-step tasks are numbered, tangents are suppressed, and lists are intentionally kept short. The style also restates the current state, makes completed progress visible, and describes errors without unnecessary drama. Every response should end with one concrete next step rather than a generic invitation or closing phrase. The skill can be installed as a plugin for Claude Code and Codex, with guidance available for other coding agents. Because its behavior is defined in an editable skill file, users can fork and tune the rules for their own workflow.

Hallmark is an anti-AI-slop design skill for AI coding assistants such as Claude Code, Cursor, and Codex. It gives agents a concrete design rule set for producing interfaces that look intentionally crafted rather than generic, repetitive, or template-like. The project focuses on typography, color, layout, motion, microinteractions, visual hierarchy, and structural variation. It can be used to build new UI, audit existing code, or redesign an interface while keeping the underlying content and information architecture. ...