agent-skills vs context-engineering-kit

AI coding-agent workflows compared

agent-skills and context-engineering-kit provide skills and workflows to guide AI coding agents through development tasks. agent-skills emphasizes portable lifecycle workflows, while context-engineering-kit focuses on selective plugins, subagent coordination, and reflection.

agent-skillscontext-engineering-kit
LanguageJavaScriptTypeScript
LicenseMITGPL-3.0
Stars109k1.8k
Forks11.3k163
Last analyzedOct 20, 2026Oct 25, 2026

Key differences

  • agent-skills provides 25 skills spanning requirements, planning, implementation, testing, security, and deployment; context-engineering-kit centers on focused plugins for specification-driven work, subagents, review, and reflection.
  • agent-skills includes nine lifecycle slash commands for Claude Code; context-engineering-kit lets Claude Code users install individual plugins and save lessons with `/memorize`.
  • agent-skills is written in JavaScript and licensed under MIT; context-engineering-kit is written in TypeScript and licensed under GPL-3.0.
  • Both support multiple agent clients, but context-engineering-kit notes that feature coverage varies by client, including reduced subagent support in some installation paths.
  • agent-skills lists 109k stars and 11.3k forks, while context-engineering-kit lists 1.8k stars and 163 forks.
  • agent-skills describes a contributor base not concentrated around one person; context-engineering-kit reports contribution activity heavily concentrated in one contributor.

Choose agent-skills if you…

  • need lifecycle workflows covering scoping through shipping across supported agent tools.
  • want explicit exit criteria, evidence requirements, and quality checklists for agent work.
  • prefer an MIT-licensed project with a contributor base that is not concentrated around one person.
Read the agent-skills analysis →

Choose context-engineering-kit if you…

  • want to install selected plugins and keep unrelated skills and agents out of the context.
  • need subagent workflows for parallel or sequential implementation with review gates.
  • want reflection and critique loops that can preserve reusable lessons with `/memorize`.
Read the context-engineering-kit analysis →

Compare the numbers side by side, or add more projects

This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.