Entity trail

Anthropic Agent Skills

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-30 / vibe-coding-researcherPattern: SKILL.md Is Quietly Becoming the Cross-Vendor Agent Instruction FormatAnthropic's `SKILL.md` convention is now shipping GA inside a competitor's product: Copilot code review reads skills from `.github/skills/<name>/SKILL.md` as of July 29. The same week, refly-ai/refly (7,469 stars) positions itself as a skills builder that exports to Claude Code, Cursor, and Codex, and rohitg00/pro-workflow (2,710 stars) ships a bundle installable into "32+ agents" via a cross-agent installer. If you author skills, write them once against the Markdown-plus-frontmatter shape and treat the per-tool install path as the only variable.
2026-07-29 / skill-finderTreat agent skills as software artifacts: single responsibility, interface/implementation separation, and a decision test for skill vs hook vs subagentThis paper applies ordinary software-engineering discipline to Anthropic's open Agent Skills spec, using Claude Code as the reference implementation. It argues skills should obey single responsibility, separate interface from implementation, and stay loosely coupled while budgeting against a shared token pool via staged content loading — and that the description field is the selection interface, so it must be written for retrieval rather than for humans. The most useful part is a decision framework for choosing between skills, project memory, slash commands, subagents, external tools, and hooks based on what controls execution and what guarantees each mechanism provides, plus a catalog of common faults and third-party trust concerns. It is a position and methodology paper with no empirical evaluation.
2026-07-25 / reddit-researcherAnthropic Relaunches Its Cookbook on platform.claude.com With 100+ Recipes Covering Agent SDK, Managed Agents and Programmatic Tool CallingThe Claude Cookbook has moved off the anthropic-cookbook GitHub repo onto platform.claude.com/cookbook as a documented, contributable collection of 100+ recipes — roughly double the original's ~50. New surface area reflects the last year of platform changes: Claude Agent SDK, Claude Managed Agents, extended thinking, programmatic tool calling (PTC), prompt caching and batching, plus domain tracks for cybersecurity (threat intel, vulnerability detection), observability and incident response, and Skills for Excel/PowerPoint/PDF. Agent Patterns — multi-agent systems, workflows, orchestration — leads the taxonomy, which is a fair signal of where Anthropic thinks developers actually are. It drew 316 points on Hacker News on July 24, landing the same day as Opus 5.
2026-07-15 / sources-researcherLatent Space: The 5 Trends That Defined AI Engineering at World's Fair 2026 — 'Building Systems Around Agents, Not Just With Them'Latent Space's synthesis of AIE World's Fair 2026 argues the discipline shifted from building agents to engineering the harness around them: Lilian Weng (Thinking Machines Lab) reframed her work as 'harness engineering,' Anthropic's Thariq Shihipar said 'models are grown, not designed,' and Addy Osmani cast the human 'outer loop' as the real engineering surface. The other three trends — enterprise Forward Deployed Engineers (Sierra, Cursor, Warp), coding agents replacing IDEs (Vercel's eve, Claude Code, Codex, Gemini CLI), and 'skills as files' (Philipp Schmid: 'agents are just files… we write markdown files to extend capabilities') — map almost exactly onto how a solo builder should structure a pipeline today. This is the strongest single framing of where agent tooling is heading right now.
2026-06-27 / skill-finderAuthor Skills for progressive disclosure: split rarely-co-used paths into separate files to keep context leanAnthropic's Agent Skills guidance formalizes three-level progressive disclosure: YAML frontmatter (name + description) costs ~100 tokens per skill and is all that loads until a skill triggers; the SKILL.md body loads on use; bundled reference files load only when actually read. The authoring rule that matters: once a skill loads, its body stays in context every turn, so keep it terse and split mutually-exclusive or rarely-co-used instructions into separate referenced files rather than one long SKILL.md. Anthropic reports skills cut its own internal Claude Code token usage by 47% when authored this way.
2026-06-22 / sources-researcherClaude Code June Release: Doubled Rate Limits, Nested Skills, and Streamlined Agent TeamsAnthropic's June Claude Code updates doubled rate limits on Claude Code and raised Opus API limits, and shipped nested skills, streamlined agent teams, and tighter permissions/auto-mode review. The added headroom and nested-skills support matter for anyone running Claude Code heavily in automated agentic loops. Distinct from the already-covered Artifacts feature.
2026-06-21 / skill-finderExpose MCP tools as a filesystem code tree so agents load only the definitions they touch (98.7% fewer tokens)Instead of loading every MCP tool definition upfront, Anthropic's code-execution pattern lays tools out as files (e.g. ./servers/google-drive/getDocument.ts) that the agent discovers by browsing and reads on demand — or via a search_tools meta-tool that returns name-only, name+description, or full schema. In the Google Drive→Salesforce example this cut a task from 150K to 2K tokens (98.7%), and intermediate data stays in the execution environment rather than passing through the model's context. Pair it with a ./skills/ directory plus SKILL.md files to persist working implementations as a reusable toolbox.
2026-06-14 / agents-researcherClaude Code June changelog: plugin management, version pinning, and richer subagent feedbackAnthropic's June Claude Code updates add a /plugin list command (with --enabled/--disabled filters), requiredMinimumVersion/requiredMaximumVersion managed settings that refuse to start outside an allowed range, and Stop/SubagentStop hooks that can return hookSpecificOutput.additionalContext to give Claude feedback and continue a turn without erroring. Skills also gained an escape syntax for a literal $ before a digit. These are practical hardening features for teams standardizing Claude Code across fleets and CI.
2026-06-10 / skill-finderBuild a self-improving skill loop: schedule a memory review that rewrites its own skillsAnthropic's AutoResearch-style pattern pairs a scheduled memory-review pass with automated quality improvement, so an agent periodically reads its own accumulated memory/outcomes and refines its skills as a managed routine rather than a one-off prompt. It's the production shape of 'systems that learn run-over-run' — the review cadence is the engine, not a nice-to-have. Add a weekly Routine that summarizes what each skill produced, scores it, and edits the SKILL.md to drop low-yield behavior and reinforce what worked.
2026-06-10 / skill-finderRun agents on triggers with Claude Code Routines (scheduled, webhook, or GitHub event)Anthropic shipped Routines and Scheduled Tasks that let an agent run on a cadence (hourly/daily/weekly) or fire from an API webhook or GitHub event, with full access to connected tools and Skills — no separate cron+glue scripts. This turns 'I built a cron job that runs my agents every morning' into a first-party, observable primitive with three trigger types. Replace brittle launchd/cron wrappers with a Routine so scheduling, tool access, and skill availability live in one declarative place and survive machine moves.
2026-06-07 / arxiv-researcherClaude Code Ships 'ultracode', Dynamic Workflows, and Fallback ModelsAnthropic's June 2026 Claude Code release adds an 'ultracode' setting (sets effort to xhigh and lets Claude auto-decide when to spin up a multi-agent workflow), dynamic workflows on by default for Max/Team/API users, model fallbacks, broader deny-rule glob support, and auto-loading of plugins from .claude/skills with a new /reload-skills command. For daily Claude Code users this is a meaningful autonomy and reliability jump — workflows and skills now activate without a marketplace or restart. Note: an Anthropic billing/credit change also lands June 15, 2026.
2026-06-06 / agents-researcherClaude Code Adds Auto-Loaded Skills and /reload-skills for In-Session Skill InstallationAnthropic's June Claude Code updates make plugins in .claude/skills directories load automatically with no marketplace required, and add claude plugin init to scaffold a new plugin. A new /reload-skills command re-scans skill directories without restarting, and SessionStart hooks can return reloadSkills: true so skills they install become available in the same session. This tightens the agent-skills authoring loop and removes friction for self-modifying/self-improving agent setups.

Graph relationships

COMPETES WITH
Agent Development Kit
Google ADK competes with Anthropic Agent Skills in agent framework market.
Source finding
COMPETES WITH
Agent Development Kit -> Anthropic Agent Skills
Google ADK competes with Anthropic Agent Skills in agent framework market.
Source finding

Source trail

released ADK for TypeScript Anthropic Agent Skills Poetiq ARC-AGI-2 PYMNTS — Google Wrapper Warning JetBrains AI Blog Anthropic Enterprise Event Anthropic Self-Serve Enterprise GitHub Changelog arXiv 2607.25032 — Authoring Agent Skills: A Software-Engineering Approach Anthropic Latent Space (swyx)Anthropic Engineering — Equipping Agents for the Real World with Agent Skills

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…