context-engineering

An agent skill by addyosmani, from addyosmani/agent-skills. Tags: developer-tools, documents.

What it does

Optimizes agent context setup. Use when starting a new session, when agent output quality degrades, when switching between tasks, or when you need to configure rules files and context for a project.

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add addyosmani/agent-skills --skill context-engineering

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/addyosmani/agent-skills
cp -r agent-skills/skills/context-engineering ~/.claude/skills/context-engineering

Safety box score

Grade B, low risk: 11 of 100 risk points, rated 2026-09-23. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This is a documentation-only skill for managing agent context: it tells the agent to read project files, create repo-local rules files, and use selective context while working. The main risks are limited to normal coding-tool activity and optional integrations that may fetch external documentation or access connected services.

Category levels, from 0 (none) to 3 (high)
Category	Level	Why
Code & shell execution	1 (low)	It references running ordinary project commands like `npm run build`, `npm test`, and `npx tsc --noEmit`, but ships no executable scripts or code-eval logic of its own.
Network access	1 (low)	The `MCP Integrations` section names external services such as `Context7`, `GitHub`, and `Chrome DevTools`, which can fetch or transmit data if the host enables them.
Filesystem reach	1 (low)	It tells the agent to create and read repo files like `CLAUDE.md`, `.cursorrules`, `.github/copilot-instructions.md`, `AGENTS.md`, and to inspect `git status`, but it does not direct edits outside the project.
Secrets & credentials	0 (none)	The only secret-related content is defensive, e.g. `Never commit .env files or secrets`; it does not ask for credentials or access key stores.
Privilege & persistence	0 (none)	There is no request for sudo, service installation, global configuration, or shell/OS changes.
Instruction hijack surface	0 (none)	The skill explicitly warns against silent assumptions with phrases like `Do NOT silently pick one interpretation` and `stop and ask`; it does not instruct the agent to ignore safeguards or hide information.
Irreversible actions	0 (none)	It does not push, deploy, send messages, or make purchases; the closest instruction is `Commit the completed task only when the user or repository workflow authorizes it`.
Transparency	0 (none)	The skill is plain Markdown with readable examples and no binaries, generated blobs, or hidden runtime-decoded content.

Before installing

Check whether you already have a repo-level rules file such as `CLAUDE.md` or `.cursorrules`; this skill is meant to add or refine one.
If your environment enables MCP servers, review which ones are connected before using the skill, especially `GitHub`, `Context7`, or `PostgreSQL`.
Confirm the repository’s policy for persistent instruction files and whether they should be committed.
Run it first in a scratch branch or non-sensitive repo to verify the context files it creates are limited to the project.

Rated 2026-09-23 by an automated review of 1 file from the skill's repository.

Source

Repository: addyosmani/agent-skills (all skills from this repository)
Path: skills/context-engineering/SKILL.md
Branch: main
Updated: 2026-09-19

code-review-and-quality — Conducts multi-axis code review. Use before merging any change. Use when reviewing code written by yourself, another agent, or a human.
constraint-driven-development — Establishes a project's quality bar as a written contract and stops agents quietly lowering it.
deprecation-and-migration — Manages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another.
documentation-and-adrs — Records decisions and documentation.
doubt-driven-development — Subjects every non-trivial decision to a fresh-context adversarial review before it stands.
git-workflow-and-versioning — Structures git workflow practices. Use when making any code change.