# Install this guide into your agent

Every section of this site is published as an Agent Skill: a folder your agent reads, carrying the architecture, the decisions and the measured evidence, with the sources intact.

Author: Murali Sid (https://linkedin.com/in/muralisid)
Source: https://www.agenticarchitectureskills.com/skills
Published: 2026-09-18
Updated: 2026-10-15
Licence: CC BY-SA 4.0, https://creativecommons.org/licenses/by-sa/4.0/
Cite as: Murali Sid. Install this guide into your agent. Agentic Architecture Skills. https://www.agenticarchitectureskills.com/skills

---


Every section of this site is published as an Agent Skill. Your agent can then use the guide while you work. An Agent Skill is a folder with a short instruction file and the reference material it needs. Agents load the instruction file only when it is relevant. So a large library costs almost nothing until it is used. You install the material once, instead of retyping pages into a chat. The skills here are reference material only. None of them asks for credentials, network access, or permission to act on your systems.

## Install everything [#install-everything]

<CopyLine command="npx skills add muralisid/agentic-architecture-skills" />

That installs the whole catalogue into whichever agents it finds. These include Claude Code, Codex, Cursor, Gemini CLI, Copilot, and anything else reading the shared `.agents/skills` location. [Per-agent instructions](/skills/install) cover the rest, including claude.ai, the Claude Code plugin, and installing by hand.

Agents that support discovery can read the index directly at [`/.well-known/agent-skills/index.json`](/.well-known/agent-skills/index.json). Every entry carries a `sha256` digest, a fingerprint of the bundle it points at. So a download can be checked before it is trusted. Installing by domain rather than by repository does not work in the current `skills` CLI. [The install page](/skills/install) has the detail.

## Task skills [#task-skills]

Each task skill is named for what an architect is trying to do. It bundles only the pages that decision needs.

<SkillsCatalog kind="task" />

## Reference skills [#reference-skills]

Each reference skill holds a whole section of the site in one bundle. Use one when the question is not known in advance.

<SkillsCatalog kind="reference" />

## What is inside a bundle [#what-is-inside-a-bundle]

Each bundle holds a short instruction file, the pages it draws on as plain Markdown, and sometimes a script.

* `SKILL.md` is the procedure, under 500 lines.
* `references/` holds the pages it needs, converted to plain Markdown. Figures are described in words and charts become tables. Every link points back at the page it came from.
* `scripts/` holds work that should be arithmetic rather than judgement. Only some bundles have one.

The readiness skill's scorer is the current example of a script. It reproduces the scoring rules of the published [readiness assessment](/library/frameworks/readiness-assessments). So two people scoring the same answers get the same profile.

Bundles are byte-reproducible. The same content always produces the same digest. So a digest that changes means the content changed, not the build.

## What they will not do [#what-they-will-not-do]

* Act on your systems. There are no tools, no network calls, and no credentials in any bundle.
* Replace the pages. The material is the same. Every reference names the page it came from. So an agent can cite the source rather than paraphrase it.
* Stay silently stale. Each bundle carries the date of its newest input. The pages behind them carry their own evidence status and re-verification dates.

## If a skill is missing [#if-a-skill-is-missing]

The catalogue grows with the site. A skill exists when there are pages behind it worth bundling, so the gaps are deliberate. If you want one that is not here, [open an issue](https://github.com/muralisid/agentic-architecture-skills) on the skills repository. Say what you were trying to do.

<details>
  <summary>
    Explore this topic
  </summary>

  * [How the skills are built](/skills/how-skills-are-built), including the conversion, the reproducibility and the checks
  * [The agent platform layer (R07)](/layers/r07-agent-platform), on skills as a capability format and what the evidence says about their limits
</details>