Description
context-engineering is an agent skill from Addy Osmani's agent-skills collection for feeding an AI agent the right information at the right time. The skill treats context as the biggest lever on agent output quality: too little context and the agent hallucinates, too much and it loses focus. Context engineering is the practice of deliberately choosing what the agent sees, when it sees it, and how that material is structured. The skill applies when starting a new coding session, when output quality slips with wrong patterns, hallucinated APIs or ignored conventions, when switching between parts of a codebase, when setting up a new project for AI-assisted development, and when the agent is not following project conventions. Its main organizing idea is the context hierarchy: structure context from the most persistent to the most transient, so that long-lived project conventions sit above the details of the current task. The skill is listed on skills.sh, with its security audit results shown there. It suits developers who work with coding agents over long sessions or across several parts of a codebase and want consistent results.…
Highlights
- Chooses what context the agent sees and when
- Context hierarchy from persistent to transient
- Fixes hallucinated APIs and ignored conventions
- Guidance for new projects and new sessions
Ratings & Reviews
No ratings yet
Be the first to rate context-engineering.
People also visit
coding-agent
Delegate coding work to Codex, Claude Code, or OpenCode as background workers
recursive-decision-ledger
Run repeated rollouts ("Prime Gauss" style recursive prompting) while keeping an
webapp-testing
Test local web applications using Playwright for UI verification and debugging
content-research-writer
Assists in writing high-quality content by conducting research
batch-grill-me
A relentless interview that asks every frontier question at once, round by round
context-engineering FAQ
What does the context-engineering skill do?
It helps structure and curate the information a coding agent sees, so its output stays accurate and follows the project's conventions.
How do I install it?
Run: npx skills add https://github.com/addyosmani/agent-skills --skill context-engineering
What is the context hierarchy?
Structuring context from the most persistent material to the most transient, so long-lived conventions sit above the details of the current task.
When should I use it?
At the start of a session, when output quality declines, when switching parts of a codebase, or when the agent ignores conventions.
What security checks has it passed?
skills.sh lists three audits: Gen Agent Trust Hub Pass, Socket Pass and Snyk Warn. A Warn or Fail is worth reading before you install.
What are the best context-engineering alternatives?
Popular alternatives to context-engineering include coding-agent, recursive-decision-ledger and webapp-testing. See all context-engineering alternatives