# Agent Skills Explained: The SKILL.md Open Standard > What Agent Skills are, the exact SKILL.md field constraints, the three-level progressive-disclosure loading model, and how Skills differ from MCP tools and native function calling. Category: Reference · Updated: 2026-07-07 · Tags: agent-skills, skill-md, standard, agents, context-engineering, progressive-disclosure, anthropic Canonical: https://changegamer.ai/resources/agent-skills-explained Variants: [HTML](https://changegamer.ai/resources/agent-skills-explained) · [JSON](https://changegamer.ai/api/resources/agent-skills-explained.json) License: https://changegamer.ai/license.xml · Access: free Agent Skills is an open, filesystem-based format for packaging procedural knowledge — reusable instructions, scripts, and reference material — that agentic coding tools load into an LLM's context only when needed. Anthropic introduced it in December 2025 as a specification independent of Claude's own implementation, published separately at agentskills.io. ## Key facts - A Skill is a directory containing a required `SKILL.md` file plus optional bundled files (scripts/, reference docs, templates). - `SKILL.md` needs only two YAML frontmatter fields: `name` (max 64 chars, lowercase letters/numbers/hyphens only, no XML tags, cannot contain "anthropic" or "claude") and `description` (non-empty, max 1024 chars, no XML tags, must state both what the skill does and when to use it). - Skills load in three stages ("progressive disclosure"): Level 1 metadata (name + description, ~100 tokens, always in the system prompt); Level 2 the SKILL.md body (read only when triggered, recommended under 5,000 tokens); Level 3 bundled scripts/resources (read or executed only as referenced — effectively unlimited). - Bundled scripts run via bash/code execution and return only their output to context — the script source itself never consumes tokens. - On the Claude API, Skills require three beta headers: `code-execution-2025-08-25`, `skills-2025-10-02`, `files-api-2025-04-14`. - Example Skills are Apache-2.0 licensed; Anthropic's own pre-built document Skills (docx/pdf/pptx/xlsx) are source-available, not open source. - Multiple independent tools now read the same format: Claude Code, the Claude API, and OpenAI's Codex CLI (which publishes its own Agent Skills docs) are directly corroborated; community trackers additionally list Google's Gemini CLI and further integrations (GitHub Copilot, Cursor, Cline, Windsurf) — exact adoption counts vary by source and were not all independently verified here. ## How Skills differ from MCP and function calling Skills are not a network protocol — there is no JSON-RPC handshake or `tools/list` call. A host discovers Skills by loading directory metadata into the system prompt at startup, then reads the full file from the filesystem via bash only when relevant. See /resources/mcp-vs-function-calling for how MCP and native function/tool calling differ from each other. The three layers compose: a Skill's instructions can tell an agent to call an MCP tool or invoke a function — Skills package *how* to do a task, MCP and function calling provide *what* the agent can call. ## Where Skills run - **Claude API** — no network access inside the execution container, only pre-installed packages; reference pre-built Skills by `skill_id` (pptx, xlsx, docx, pdf) or upload custom ones via `/v1/skills` (workspace-shared). - **Claude Code** — filesystem-based, no upload step; personal (`~/.claude/skills/`) or project-scoped (`.claude/skills/`); full local network access. - **claude.ai** — custom Skills are per-user (not org-wide), uploaded as a zip via Settings; network access depends on admin configuration. ## Security note A Skill runs with whatever tool/file/network access the host already grants the agent — a malicious or compromised Skill (especially one bundling code that fetches external content) can exfiltrate data or misuse tools exactly like a hostile MCP server or a poisoned prompt. Audit every bundled file before installing a Skill from an untrusted source. See /resources/agentic-security-checklist for the general threat model and /resources/agent-memory-context for how "procedural memory" as a concept relates to this concrete format. ## Verified sources - Agent Skills overview (Claude Platform docs, fetched directly): https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview - anthropics/skills repository README (fetched directly): https://github.com/anthropics/skills - Agent Skills open standard specification (referenced by the above; not independently fetched — 403 in this sandbox): https://agentskills.io/specification - Anthropic engineering blog, "Equipping agents for the real world with Agent Skills" (cited by the docs above; not independently fetched — anthropic.com is sandbox-blocked): https://www.anthropic.com/engineering/equipping-agents-for-the-real-world-with-agent-skills - OpenAI Codex — Agent Skills docs (existence and title corroborated via search index; not independently fetched — 403 in this sandbox): https://developers.openai.com/codex/skills --- ## Related resources - [AGENTS.md Explained: The Open Standard for Repo-Level Agent Instructions](https://changegamer.ai/resources/agents-md-explained.md): What AGENTS.md is, why OpenAI created it, how it differs from SKILL.md and a human-facing README, which coding agents read it today, and what this session could and could not independently confirm about its move to the Linux Foundation. - [C2PA Content Credentials: Verifying Media Provenance and AI-Generation Claims](https://changegamer.ai/resources/c2pa-content-credentials.md): How the C2PA standard cryptographically signs images, video, and audio with provenance manifests recording capture, edit, and AI-generation history — what a manifest contains, how a verifier checks one, and why a missing manifest proves nothing either way. - [Content Signals Explained: The robots.txt Extension for AI Usage Intent](https://changegamer.ai/resources/content-signals-explained.md): What Cloudflare's Content Signals Policy and the IETF AIPREF draft add to robots.txt — three usage-intent directives (search, ai-input, ai-train) that declare preferences, not access control, and how they differ from crawler-blocking tokens and RSL. - [LLM Model Deprecation: Detecting and Handling End-of-Life Models](https://changegamer.ai/resources/llm-model-deprecation.md): How the general IETF Sunset/Deprecation HTTP headers work, why none of the three major LLM APIs actually send them, and each vendor's real notice periods and retirement mechanics — plus detection and fallback patterns for agents that pin a model ID today. --- Index of all resources: https://changegamer.ai/llms.txt · Full corpus: https://changegamer.ai/llms-full.txt · Corpus data (NDJSON): https://changegamer.ai/api/corpus.jsonl · Offers: https://changegamer.ai/api/pricing.json