标签 · Agent Skills — Glean

Recent picks

7picks · chronological

08-31

The new rules of context engineering for Claude 5 generation models

Thariq Shihipar, a member of technical staff at Anthropic, explains how context engineering for Claude Code has changed with Claude 5-generation models. The old playbook—hard rules, examples, everything upfront, repeated instructions, CLAUDE.md as memory, and simple markdown specs—is now partly myth. New guidance: let the model use judgement (the system prompt now says 'match the surrounding code' instead of banning multiline comments); design expressive tool interfaces instead of stuffing examples; adopt progressive disclosure by loading verification and review skills on demand and letting tools defer their definitions until ToolSearch finds them. CLAUDE.md should stay lightweight, holding repo-specific gotchas, while skills encode team-specific opinions. Specs can become rich references: HTML mockups, test suites, code to port, or rubrics with verifier agents. Anthropic also ships `claude doctor` to auto-simplify contexts. No benchmarks are included, and the article is vendor-authored, but it offers immediately actionable patterns for agent harness and Claude Code users.

claude.com · 7 min · Agent Engineering · Agent Skills · Claude Code

08-27

An agent skill that turns codebases into verifiable, interactive system maps

Archify is an agent skill that turns a system description or an entire repository into an interactive technical map. It supports five diagram types—architecture, workflow, sequence, data-flow, and lifecycle—and produces a single self-contained HTML file with built-in navigation, dark/light themes, finite motion, and exports to PNG, SVG, WebM, and 1200×630 share cards. Under the hood it uses a typed JSON IR as the source of truth, runs validation before delivery, and returns machine-readable repair receipts on failure. It also supports Before/Delta/After snapshots for architecture review. Designed for Raven, Cursor, Claude Code, Codex CLI, and OpenCode, it suits engineers and doc writers who want verifiable, shareable visualizations without hand-drawing.

github.com · 15 min · Agent Skills · AI Engineering · CLI

08-24

At AIEWF, AI Engineer Trends Shift from Autonomy to Harnessed Loops

At the 2026 AI Engineer World’s Fair, the conversation has moved from autonomous agents to harnesses, loops, and skills. Lilian Weng's new essay shifts from planning/memory/tool-use anatomy to harness engineering around the model. AutoGPT-era autonomy is out; engineers like OpenClaw's Peter Steinberger put themselves in an outer loop while agents run the inner loop. Enterprises are adopting forward deployed engineers who integrate long-running agents and software factories (Cursor, Warp, Vercel) with strict ROI targets. Google DeepMind's Philipp Schmid says "agents are just files" and Markdown-based skills extend capabilities; Matt Pocock warns of "skills hell" and advises fewer, smaller skills. Skepticism remains: HumanLayer's Dex Horthy says hype outruns discipline, and Geoffrey Huntley worries that "factories failed, loops failed" will be next year's refrain. A useful orientation for engineers tracking AI development tools and agentic workflows.

www.latent.space · 19 min · Agent Engineering · Agent Skills · Coding Agents

08-20

A shared browser for AI agents: parallel automation, shared logins

ego lite is a macOS browser purpose-built for AI agents such as Claude Code and Codex. It lets agents run browser automation in parallel inside isolated Spaces while you continue browsing in the same browser — no tab wars, no interrupted flow. Unlike frameworks like Browser-use that drive a separate browser, ego lite is a full browser that can inherit your existing Chrome logins, cookies, extensions, and bookmarks in one click. It exposes browser capabilities as in-page JavaScript tools via the `ego-browser` skill, so agents can compose multi-step actions as single code executions instead of round-trip CLI calls. The README reports up to 2.5× faster task completion and fewer tokens than a CLI-based approach on complex tasks, with data kept locally by default. For engineers using agent CLIs today, ego lite offers a straightforward way to give agents safe, stateful access to the web while keeping your own session separate.

github.com · 10 min · Agent Skills · AI Agents · Browser Automation

08-20

An editorial diagram design system for AI coding agents

Diagram Design is a skill package for AI coding agents (Claude Code, Codex, Factory Droid, and Pi) that produces editorial-grade, self-contained HTML+SVG diagrams. It tackles the generic, cheap-looking output that agents usually generate by codifying a tight design system — one accent color, hairline borders, strict 4px geometry, no shadows, and three static variants across 38 diagram types. A 60-second onboarding flow pulls colors and fonts from your website into semantic tokens; import commands redraw draw.io or Mermaid sources with adjustable format, size, detail, and audience; accessibility checks (WCAG AA contrast, accessible SVG names) are enforced by default. For engineers and writers who want AI-generated diagrams that don't need manual cleanup.

github.com · 40 min · Agent Skills · Claude Code · Data Visualization

07-10

Lessons from Building Claude Code: How We Use Skills

Anthropic engineers share practical lessons from maintaining hundreds of Claude Code skills internally. The post categorizes skills into nine types—library reference, product verification, data fetching, business process automation, code scaffolding, code quality, CI/CD, runbooks, and infrastructure ops—and offers concrete writing tips: focus on non-obvious hints, build gotchas sections, use the filesystem for progressive disclosure, avoid over-constraint, design setup flows, treat the description field as a trigger for the model, store memory (logs, JSON, SQLite), and bundle scripts/libraries to reduce token waste. It also covers distribution strategies (in-repo vs. marketplace), skill composition, and usage measurement via PreToolUse hooks. A must-read for anyone building agent-centric workflows.

x.com · 15 min · Agent Engineering · Agent Skills · Anthropic

06-30

5 Agent Skills I Use Every Day to Encode My Development Process

Matt Pocock, a seasoned engineer, shares 5 agent skills he uses daily to encode rigorous, repeatable processes for LLM agents, addressing their lack of memory and tendency to drift. Key skills include: grill-me (exhaustive questioning before coding), to-prd (turning discussions into PRDs), to-issues (slicing PRDs into vertical issues), tdd (forcing red-green-refactor cycles for quality), and improve-codebase-architecture (identifying shallow modules for deepening). The core insight: short, well-crafted skills can dramatically boost agent output quality.

www.aihero.dev · 8 min · Agent Skills · Claude Code · Process Encoding