X postHamel HusainAnnouncement
Evaluating AI outputs and traces manually is time-consuming. Equipping coding agents with skills to turn raw execution traces into review interfaces streamlines debugging and assessment workflows.
The update focuses on bridging the gap between raw AI trace logs and actionable inspection. Leveraging coding agents to spin up purpose-built review interfaces reduces the overhead required to inspect system failures and refine model evaluations.
Hamel Husain
@HamelHusain · 17 August 2026
A few months ago,@sh_reya and I released eval skills plugin. We iterated on it a bunch and recently made it better The biggest change is a new error-discovery skill. Give your coding agent a file of AI outputs or traces, and it builds a custom review app w/intelligent
717 likes41 replies1.1k bookmarks71k views
View on XSummary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.
Similar in content, found by the AI search.
Repository:EveryInc/compound-engineering-plugin
Compound Engineering is an open-source plugin for AI coding agents designed to structure development workflows and systematically preserve learned context. It runs on 14 platforms, including Claude Code, Cursor, and Codex.
24.9kstarsTypeScript
X post:Emre Savcı
Cloudflare has open-sourced security-audit-skill, a coding-agent skill designed to run multi-phase, automated security audits on codebases. It coordinates isolated agent roles across reconnaissance, vulnerability hunting, and independent verification.
2.8kbookmarks143.6kviews
X post:Adham Dannaway
Adham Dannaway highlights 'Design Skills for AI Agents', an open collection developed by Carlos Cuéllar. The repository provides structured design heuristics, WCAG 2.2 accessibility rules, and design system frameworks for tools such as Claude Code, Cursor, and Codex.
100bookmarks5.2kviews
Repository:petergyang/human-review
Human Review is an open-source tool enabling users to visually edit and annotate HTML files, Markdown documents, and localhost web pages directly in the browser. Instead of writing tedious chat prompts describing UI tweaks, it captures direct text edits and Google Docs-style comments in a single batch and returns them to coding agents like Claude Code or Codex.
1.2kstarsJavaScript
X post:Zach Lloyd
Zach Lloyd shares that Warp's software factory has automatically generated and merged ten skill improvement PRs. This workflow is powered by Warp Factories' new self-improvement feature, which monitors agent runs for inefficiencies and autonomously proposes fixes to agent prompts and skill configurations.
298bookmarks24.3kviews
Link:cuellarfr
cuellarfr/design-skills provides ten structured skill packages for developer agents including Claude Code, Cursor, Codex, and OpenCode. The library covers areas ranging from UX research and accessibility to design systems and UX writing.
Parts of this content were created with the assistance of AI.
